modelpedia 0.0.3 → 0.0.5

package/dist/providers/openrouter.cjs

@@ -1,12 +1,21 @@
-Object.defineProperty(exports,Symbol.toStringTag,{value:`Module`});const e={id:`openrouter`,name:`OpenRouter`,region:`US`,url:`https://openrouter.ai`,api_url:`https://openrouter.ai/api/v1`,docs_url:`https://openrouter.ai/docs`,pricing_url:`https://openrouter.ai/models`,icon:`<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="currentColor" fill-rule="evenodd">
+Object.defineProperty(exports,Symbol.toStringTag,{value:`Module`});const e={id:`openrouter`,name:`OpenRouter`,region:`US`,headquarters:`New York, NY`,founded:2023,url:`https://openrouter.ai`,api_url:`https://openrouter.ai/api/v1`,docs_url:`https://openrouter.ai/docs`,pricing_url:`https://openrouter.ai/models`,description:`Unified API gateway providing access to 200+ models from multiple providers with standardized pricing.`,type:`aggregator`,playground_url:`https://openrouter.ai/playground`,status_url:`https://status.openrouter.ai`,openai_compatible:!0,free_tier:!0,github_url:`https://github.com/OpenRouterTeam`,models_url:`https://openrouter.ai/models`,twitter_url:`https://x.com/OpenRouterAI`,discord_url:`https://discord.com/invite/openrouter`,terms_url:`https://openrouter.ai/terms`,support_url:`https://openrouter.ai/support`,icon:`<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="currentColor" fill-rule="evenodd">
   <path d="M16.804 1.957l7.22 4.105v.087L16.73 10.21l.017-2.117-.821-.03c-1.059-.028-1.611.002-2.268.11-1.064.175-2.038.577-3.147 1.352L8.345 11.03c-.284.195-.495.336-.68.455l-.515.322-.397.234.385.23.53.338c.476.314 1.17.796 2.701 1.866 1.11.775 2.083 1.177 3.147 1.352l.3.045c.694.091 1.375.094 2.825.033l.022-2.159 7.22 4.105v.087L16.589 22l.014-1.862-.635.022c-1.386.042-2.137.002-3.138-.162-1.694-.28-3.26-.926-4.881-2.059l-2.158-1.5a21.997 21.997 0 00-.755-.498l-.467-.28a55.927 55.927 0 00-.76-.43C2.908 14.73.563 14.116 0 14.116V9.888l.14.004c.564-.007 2.91-.622 3.809-1.124l1.016-.58.438-.274c.428-.28 1.072-.726 2.686-1.853 1.621-1.133 3.186-1.78 4.881-2.059 1.152-.19 1.974-.213 3.814-.138l.02-1.907z"/>
-</svg>`,models:[{id:`ai21/jamba-large-1.7`,name:`AI21: Jamba Large 1.7`,created_by:`ai21`,source:`official`,last_updated:`2026-03-21`,description:`Jamba Large 1.7 is the latest model in the Jamba open family, offering improvements in grounding, instruction-following, and overall efficiency.`,context_window:256e3,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2,output:8}},{id:`aion-labs/aion-1.0-mini`,name:`AionLabs: Aion-1.0-Mini`,created_by:`aion-labs`,source:`official`,last_updated:`2026-03-21`,description:`Aion-1.0-Mini 32B parameter model is a distilled version of the DeepSeek-R1 model, designed for strong performance in reasoning domains such as mathematics, coding, and logic.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:.7,output:1.4}},{id:`aion-labs/aion-1.0`,name:`AionLabs: Aion-1.0`,created_by:`aion-labs`,source:`official`,last_updated:`2026-03-21`,description:`Aion-1.0 is a multi-model system designed for high performance across various tasks, including reasoning and coding.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:4,output:8}},{id:`aion-labs/aion-2.0`,name:`AionLabs: Aion-2.0`,created_by:`aion-labs`,source:`official`,last_updated:`2026-03-21`,description:`Aion-2.0 is a variant of DeepSeek V3.2 optimized for immersive roleplaying and storytelling.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:.8,output:1.6,cached_input:.2}},{id:`aion-labs/aion-rp-llama-3.1-8b`,name:`AionLabs: Aion-RP 1.0 (8B)`,created_by:`aion-labs`,source:`official`,last_updated:`2026-03-21`,description:`Aion-RP-Llama-3.1-8B ranks the highest in the character evaluation portion of the RPBench-Auto benchmark, a roleplaying-specific variant of Arena-Hard-Auto, where LLMs evaluate each other’s responses.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.8,output:1.6}},{id:`alfredpros/codellama-7b-instruct-solidity`,name:`AlfredPros: CodeLLaMa 7B Instruct Solidity`,created_by:`alfredpros`,source:`official`,last_updated:`2026-03-21`,description:`A finetuned 7 billion parameters Code LLaMA - Instruct model to generate Solidity smart contract using 4-bit QLoRA finetuning provided by PEFT library.`,context_window:4096,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},pricing:{input:.8,output:1.2}},{id:`alibaba/tongyi-deepresearch-30b-a3b`,name:`Tongyi DeepResearch 30B A3B`,created_by:`alibaba`,source:`official`,last_updated:`2026-03-21`,description:`Tongyi DeepResearch is an agentic large language model developed by Tongyi Lab, with 30 billion total parameters activating only 3 billion per token.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.09,output:.45,cached_input:.09}},{id:`allenai/molmo-2-8b`,name:`AllenAI: Molmo2 8B`,created_by:`allenai`,source:`official`,last_updated:`2026-03-21`,description:`Molmo2-8B is an open vision-language model developed by the Allen Institute for AI (Ai2) as part of the Molmo2 family, supporting image, video, and multi-image understanding and grounding.`,status:`deprecated`,deprecation_date:`2026-03-23`,context_window:36864,max_output_tokens:36864,modalities:{input:[`text`,`image`,`video`],output:[`text`]},pricing:{input:.2,output:.2}},{id:`allenai/olmo-2-0325-32b-instruct`,name:`AllenAI: Olmo 2 32B Instruct`,created_by:`allenai`,source:`official`,last_updated:`2026-03-21`,description:`OLMo-2 32B Instruct is a supervised instruction-finetuned variant of the OLMo-2 32B March 2025 base model.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:.05,output:.2}},{id:`allenai/olmo-3-32b-think`,name:`AllenAI: Olmo 3 32B Think`,created_by:`allenai`,source:`official`,last_updated:`2026-03-21`,description:`Olmo 3 32B Think is a large-scale, 32-billion-parameter model purpose-built for deep reasoning, complex logic chains and advanced instruction-following scenarios.`,context_window:65536,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.15,output:.5}},{id:`allenai/olmo-3-7b-instruct`,name:`AllenAI: Olmo 3 7B Instruct`,created_by:`allenai`,source:`official`,last_updated:`2026-03-21`,description:`Olmo 3 7B Instruct is a supervised instruction-fine-tuned variant of the Olmo 3 7B base model, optimized for instruction-following, question-answering, and natural conversational dialogue.`,status:`deprecated`,deprecation_date:`2026-03-23`,context_window:65536,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.1,output:.2}},{id:`allenai/olmo-3-7b-think`,name:`AllenAI: Olmo 3 7B Think`,created_by:`allenai`,source:`official`,last_updated:`2026-03-21`,description:`Olmo 3 7B Think is a research-oriented language model in the Olmo family designed for advanced reasoning and instruction-driven tasks.`,status:`deprecated`,deprecation_date:`2026-03-23`,context_window:65536,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.12,output:.2}},{id:`allenai/olmo-3.1-32b-instruct`,name:`AllenAI: Olmo 3.1 32B Instruct`,created_by:`allenai`,source:`official`,last_updated:`2026-03-21`,description:`Olmo 3.1 32B Instruct is a large-scale, 32-billion-parameter instruction-tuned language model engineered for high-performance conversational AI, multi-turn dialogue, and practical instruction following.`,context_window:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.2,output:.6}},{id:`allenai/olmo-3.1-32b-think`,name:`AllenAI: Olmo 3.1 32B Think`,created_by:`allenai`,source:`official`,last_updated:`2026-03-21`,description:`Olmo 3.1 32B Think is a large-scale, 32-billion-parameter model designed for deep reasoning, complex multi-step logic, and advanced instruction following.`,context_window:65536,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.15,output:.5}},{id:`alpindale/goliath-120b`,name:`Goliath 120B`,created_by:`alpindale`,source:`official`,last_updated:`2026-03-21`,description:`A large LLM created by combining two fine-tuned Llama 70B models into one 120B model.`,context_window:6144,max_output_tokens:1024,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:3.75,output:7.5}},{id:`amazon/nova-2-lite-v1`,name:`Amazon: Nova 2 Lite`,created_by:`amazon`,source:`official`,last_updated:`2026-03-21`,description:`Nova 2 Lite is a fast, cost-effective reasoning model for everyday workloads that can process text, images, and videos to generate text.`,context_window:1e6,max_output_tokens:65535,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:.3,output:2.5}},{id:`amazon/nova-lite-v1`,name:`Amazon: Nova Lite 1.0`,created_by:`amazon`,source:`official`,last_updated:`2026-03-21`,description:`Amazon Nova Lite 1.0 is a very low-cost multimodal model from Amazon that focused on fast processing of image, video, and text inputs to generate text output.`,context_window:3e5,max_output_tokens:5120,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.06,output:.24}},{id:`amazon/nova-micro-v1`,name:`Amazon: Nova Micro 1.0`,created_by:`amazon`,source:`official`,last_updated:`2026-03-21`,description:`Amazon Nova Micro 1.0 is a text-only model that delivers the lowest latency responses in the Amazon Nova family of models at a very low cost.`,context_window:128e3,max_output_tokens:5120,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.035,output:.14}},{id:`amazon/nova-premier-v1`,name:`Amazon: Nova Premier 1.0`,created_by:`amazon`,source:`official`,last_updated:`2026-03-21`,description:`Amazon Nova Premier is the most capable of Amazon’s multimodal models for complex reasoning tasks and for use as the best teacher for distilling custom models.`,context_window:1e6,max_output_tokens:32e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:2.5,output:12.5,cached_input:.625}},{id:`amazon/nova-pro-v1`,name:`Amazon: Nova Pro 1.0`,created_by:`amazon`,source:`official`,last_updated:`2026-03-21`,description:`Amazon Nova Pro 1.0 is a capable multimodal model from Amazon focused on providing a combination of accuracy, speed, and cost for a wide range of tasks.`,context_window:3e5,max_output_tokens:5120,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.8,output:3.2}},{id:`anthracite-org/magnum-v4-72b`,name:`Magnum v4 72B`,created_by:`anthracite-org`,source:`official`,last_updated:`2026-03-21`,description:`This is a series of models designed to replicate the prose quality of the Claude 3 models, specifically Sonnet(https://openrouter.ai/anthropic/claude-3.5-sonnet) and Opus(https://openrouter.ai/anthropic/claude-3-opus).`,context_window:16384,max_output_tokens:2048,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:3,output:5}},{id:`anthropic/claude-3-haiku`,name:`Anthropic: Claude 3 Haiku`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-21`,family:`claude-haiku`,description:`Claude 3 Haiku is Anthropic's fastest and most compact model for
+</svg>`,models:[{id:`ai21/jamba-large-1.7`,name:`AI21: Jamba Large 1.7`,created_by:`ai21`,source:`official`,last_updated:`2026-03-23`,description:`Jamba Large 1.7 is the latest model in the Jamba open family, offering improvements in grounding, instruction-following, and overall efficiency.`,context_window:256e3,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2,output:8},model_type:`chat`,release_date:`2025-08-08`,tagline:`Jamba Large 1.7 is the latest model in the Jamba open family, offering improvements in grounding, instruction-following, and overall efficiency.`,tools:[`function_calling`]},{id:`aion-labs/aion-1.0-mini`,name:`AionLabs: Aion-1.0-Mini`,created_by:`aion-labs`,source:`official`,last_updated:`2026-03-23`,description:`Aion-1.0-Mini 32B parameter model is a distilled version of the DeepSeek-R1 model, designed for strong performance in reasoning domains such as mathematics, coding, and logic.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:.7,output:1.4},release_date:`2025-02-04`,tagline:`Aion-1.0-Mini 32B parameter model is a distilled version of the DeepSeek-R1 model, designed for strong performance in reasoning domains such as mathematics, coding, and logic.`},{id:`aion-labs/aion-1.0`,name:`AionLabs: Aion-1.0`,created_by:`aion-labs`,source:`official`,last_updated:`2026-03-23`,description:`Aion-1.0 is a multi-model system designed for high performance across various tasks, including reasoning and coding.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:4,output:8},release_date:`2025-02-04`,tagline:`Aion-1.0 is a multi-model system designed for high performance across various tasks, including reasoning and coding.`},{id:`aion-labs/aion-2.0`,name:`AionLabs: Aion-2.0`,created_by:`aion-labs`,source:`official`,last_updated:`2026-03-23`,description:`Aion-2.0 is a variant of DeepSeek V3.2 optimized for immersive roleplaying and storytelling.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:.8,output:1.6,cached_input:.2},release_date:`2026-02-23`,tagline:`Aion-2.0 is a variant of DeepSeek V3.2 optimized for immersive roleplaying and storytelling.`},{id:`aion-labs/aion-rp-llama-3.1-8b`,name:`AionLabs: Aion-RP 1.0 (8B)`,created_by:`aion-labs`,source:`official`,last_updated:`2026-03-23`,description:`Aion-RP-Llama-3.1-8B ranks the highest in the character evaluation portion of the RPBench-Auto benchmark, a roleplaying-specific variant of Arena-Hard-Auto, where LLMs evaluate each other’s responses.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.8,output:1.6},parameters:8,release_date:`2025-02-04`,tagline:`Aion-RP-Llama-3.1-8B ranks the highest in the character evaluation portion of the RPBench-Auto benchmark, a roleplaying-specific variant of Arena-Hard-Auto, where LLMs evaluate each other’s responses.`},{id:`alfredpros/codellama-7b-instruct-solidity`,name:`AlfredPros: CodeLLaMa 7B Instruct Solidity`,created_by:`alfredpros`,source:`official`,last_updated:`2026-03-23`,description:`A finetuned 7 billion parameters Code LLaMA - Instruct model to generate Solidity smart contract using 4-bit QLoRA finetuning provided by PEFT library.`,context_window:4096,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},pricing:{input:.8,output:1.2},family:`codellama`,model_type:`code`,parameters:7,release_date:`2025-04-14`,tagline:`A finetuned 7 billion parameters Code LLaMA - Instruct model to generate Solidity smart contract using 4-bit QLoRA finetuning provided by PEFT library.`},{id:`alibaba/tongyi-deepresearch-30b-a3b`,name:`Tongyi DeepResearch 30B A3B`,created_by:`alibaba`,source:`official`,last_updated:`2026-03-23`,description:`Tongyi DeepResearch is an agentic large language model developed by Tongyi Lab, with 30 billion total parameters activating only 3 billion per token.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.09,output:.45,cached_input:.09},model_type:`chat`,parameters:30,active_parameters:3,release_date:`2025-09-18`,tagline:`Tongyi DeepResearch is an agentic large language model developed by Tongyi Lab, with 30 billion total parameters activating only 3 billion per token.`,tools:[`function_calling`]},{id:`allenai/molmo-2-8b`,name:`AllenAI: Molmo2 8B`,created_by:`allenai`,source:`official`,last_updated:`2026-03-21`,description:`Molmo2-8B is an open vision-language model developed by the Allen Institute for AI (Ai2) as part of the Molmo2 family, supporting image, video, and multi-image understanding and grounding.`,status:`deprecated`,deprecation_date:`2026-03-23`,context_window:36864,max_output_tokens:36864,modalities:{input:[`text`,`image`,`video`],output:[`text`]},pricing:{input:.2,output:.2},parameters:8},{id:`allenai/olmo-2-0325-32b-instruct`,name:`AllenAI: Olmo 2 32B Instruct`,created_by:`allenai`,source:`official`,last_updated:`2026-03-23`,description:`OLMo-2 32B Instruct is a supervised instruction-finetuned variant of the OLMo-2 32B March 2025 base model.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:.05,output:.2},parameters:32,release_date:`2025-03-14`,tagline:`OLMo-2 32B Instruct is a supervised instruction-finetuned variant of the OLMo-2 32B March 2025 base model.`},{id:`allenai/olmo-3-32b-think`,name:`AllenAI: Olmo 3 32B Think`,created_by:`allenai`,source:`official`,last_updated:`2026-03-23`,description:`Olmo 3 32B Think is a large-scale, 32-billion-parameter model purpose-built for deep reasoning, complex logic chains and advanced instruction-following scenarios.`,context_window:65536,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.15,output:.5},parameters:32,release_date:`2025-11-21`,tagline:`Olmo 3 32B Think is a large-scale, 32-billion-parameter model purpose-built for deep reasoning, complex logic chains and advanced instruction-following scenarios.`},{id:`allenai/olmo-3-7b-instruct`,name:`AllenAI: Olmo 3 7B Instruct`,created_by:`allenai`,source:`official`,last_updated:`2026-03-21`,description:`Olmo 3 7B Instruct is a supervised instruction-fine-tuned variant of the Olmo 3 7B base model, optimized for instruction-following, question-answering, and natural conversational dialogue.`,status:`deprecated`,deprecation_date:`2026-03-23`,context_window:65536,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.1,output:.2},parameters:7},{id:`allenai/olmo-3-7b-think`,name:`AllenAI: Olmo 3 7B Think`,created_by:`allenai`,source:`official`,last_updated:`2026-03-21`,description:`Olmo 3 7B Think is a research-oriented language model in the Olmo family designed for advanced reasoning and instruction-driven tasks.`,status:`deprecated`,deprecation_date:`2026-03-23`,context_window:65536,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.12,output:.2},parameters:7},{id:`allenai/olmo-3.1-32b-instruct`,name:`AllenAI: Olmo 3.1 32B Instruct`,created_by:`allenai`,source:`official`,last_updated:`2026-03-23`,description:`Olmo 3.1 32B Instruct is a large-scale, 32-billion-parameter instruction-tuned language model engineered for high-performance conversational AI, multi-turn dialogue, and practical instruction following.`,context_window:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.2,output:.6},parameters:32,release_date:`2026-01-06`,tagline:`Olmo 3.1 32B Instruct is a large-scale, 32-billion-parameter instruction-tuned language model engineered for high-performance conversational AI, multi-turn dialogue, and practical instruction followin`,tools:[`function_calling`]},{id:`allenai/olmo-3.1-32b-think`,name:`AllenAI: Olmo 3.1 32B Think`,created_by:`allenai`,source:`official`,last_updated:`2026-03-23`,description:`Olmo 3.1 32B Think is a large-scale, 32-billion-parameter model designed for deep reasoning, complex multi-step logic, and advanced instruction following.`,context_window:65536,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.15,output:.5},parameters:32,status:`deprecated`,release_date:`2025-12-16`,deprecation_date:`2026-04-06`,tagline:`Olmo 3.1 32B Think is a large-scale, 32-billion-parameter model designed for deep reasoning, complex multi-step logic, and advanced instruction following.`},{id:`alpindale/goliath-120b`,name:`Goliath 120B`,created_by:`alpindale`,source:`official`,last_updated:`2026-03-23`,description:`A large LLM created by combining two fine-tuned Llama 70B models into one 120B model.`,context_window:6144,max_output_tokens:1024,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:3.75,output:7.5},parameters:120,release_date:`2023-11-10`,tagline:`A large LLM created by combining two fine-tuned Llama 70B models into one 120B model.`},{id:`amazon/nova-2-lite-v1`,name:`Amazon: Nova 2 Lite`,created_by:`amazon`,source:`official`,last_updated:`2026-03-23`,description:`Nova 2 Lite is a fast, cost-effective reasoning model for everyday workloads that can process text, images, and videos to generate text.`,context_window:1e6,max_output_tokens:65535,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:.3,output:2.5},model_type:`chat`,release_date:`2025-12-02`,tagline:`Nova 2 Lite is a fast, cost-effective reasoning model for everyday workloads that can process text, images, and videos to generate text.`,tools:[`function_calling`]},{id:`amazon/nova-lite-v1`,name:`Amazon: Nova Lite 1.0`,created_by:`amazon`,source:`official`,last_updated:`2026-03-23`,description:`Amazon Nova Lite 1.0 is a very low-cost multimodal model from Amazon that focused on fast processing of image, video, and text inputs to generate text output.`,context_window:3e5,max_output_tokens:5120,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.06,output:.24},model_type:`chat`,release_date:`2024-12-05`,tagline:`Amazon Nova Lite 1.0 is a very low-cost multimodal model from Amazon that focused on fast processing of image, video, and text inputs to generate text output.`,tools:[`function_calling`]},{id:`amazon/nova-micro-v1`,name:`Amazon: Nova Micro 1.0`,created_by:`amazon`,source:`official`,last_updated:`2026-03-23`,description:`Amazon Nova Micro 1.0 is a text-only model that delivers the lowest latency responses in the Amazon Nova family of models at a very low cost.`,context_window:128e3,max_output_tokens:5120,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.035,output:.14},model_type:`chat`,release_date:`2024-12-05`,tagline:`Amazon Nova Micro 1.0 is a text-only model that delivers the lowest latency responses in the Amazon Nova family of models at a very low cost.`,tools:[`function_calling`]},{id:`amazon/nova-premier-v1`,name:`Amazon: Nova Premier 1.0`,created_by:`amazon`,source:`official`,last_updated:`2026-03-23`,description:`Amazon Nova Premier is the most capable of Amazon’s multimodal models for complex reasoning tasks and for use as the best teacher for distilling custom models.`,context_window:1e6,max_output_tokens:32e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:2.5,output:12.5,cached_input:.625},model_type:`chat`,release_date:`2025-10-31`,tagline:`Amazon Nova Premier is the most capable of Amazon’s multimodal models for complex reasoning tasks and for use as the best teacher for distilling custom models.`,tools:[`function_calling`]},{id:`amazon/nova-pro-v1`,name:`Amazon: Nova Pro 1.0`,created_by:`amazon`,source:`official`,last_updated:`2026-03-23`,description:`Amazon Nova Pro 1.0 is a capable multimodal model from Amazon focused on providing a combination of accuracy, speed, and cost for a wide range of tasks.`,context_window:3e5,max_output_tokens:5120,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.8,output:3.2},model_type:`chat`,release_date:`2024-12-05`,tagline:`Amazon Nova Pro 1.0 is a capable multimodal model from Amazon focused on providing a combination of accuracy, speed, and cost for a wide range of tasks.`,tools:[`function_calling`]},{id:`anthracite-org/magnum-v4-72b`,name:`Magnum v4 72B`,created_by:`anthracite-org`,source:`official`,last_updated:`2026-03-23`,description:`This is a series of models designed to replicate the prose quality of the Claude 3 models, specifically Sonnet(https://openrouter.ai/anthropic/claude-3.5-sonnet) and Opus(https://openrouter.ai/anthropic/claude-3-opus).`,context_window:16384,max_output_tokens:2048,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:3,output:5},parameters:72,release_date:`2024-10-22`,tagline:`This is a series of models designed to replicate the prose quality of the Claude 3 models, specifically Sonnet(https://openrouter.ai/anthropic/claude-3.5-sonnet) and Opus(https://openrouter.ai/anthrop`},{id:`anthropic/claude-3-haiku`,name:`Anthropic: Claude 3 Haiku`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-23`,family:`claude-haiku`,description:`Claude 3 Haiku is Anthropic's fastest and most compact model for
 near-instant responsiveness. Quick and accurate targeted performance.
 
-See the launch announcement and benchmark results [here](https:/`,context_window:2e5,max_output_tokens:4096,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.25,output:1.25,cached_input:.03}},{id:`anthropic/claude-3.5-haiku`,name:`Anthropic: Claude 3.5 Haiku`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-21`,family:`claude-haiku`,description:`Claude 3.5 Haiku features offers enhanced capabilities in speed, coding accuracy, and tool use.`,context_window:2e5,max_output_tokens:8192,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.8,output:4,cached_input:.08}},{id:`anthropic/claude-3.5-sonnet`,name:`Anthropic: Claude 3.5 Sonnet`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-21`,family:`claude-sonnet`,description:`New Claude 3.5 Sonnet delivers better-than-Opus capabilities, faster-than-Sonnet speeds, at the same Sonnet prices.`,context_window:2e5,max_output_tokens:8192,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:6,output:30,cached_input:.6}},{id:`anthropic/claude-3.7-sonnet:thinking`,name:`Anthropic: Claude 3.7 Sonnet (thinking)`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-21`,family:`claude-sonnet`,description:`Claude 3.7 Sonnet is an advanced large language model with improved reasoning, coding, and problem-solving capabilities.`,status:`deprecated`,deprecation_date:`2026-05-05`,context_window:2e5,max_output_tokens:64e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:3,output:15,cached_input:.3}},{id:`anthropic/claude-3.7-sonnet`,name:`Anthropic: Claude 3.7 Sonnet`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-21`,family:`claude-sonnet`,description:`Claude 3.7 Sonnet is an advanced large language model with improved reasoning, coding, and problem-solving capabilities.`,status:`deprecated`,deprecation_date:`2026-05-05`,context_window:2e5,max_output_tokens:64e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:3,output:15,cached_input:.3}},{id:`anthropic/claude-haiku-4.5`,name:`Anthropic: Claude Haiku 4.5`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-21`,family:`claude-haiku`,description:`Claude Haiku 4.5 is Anthropic’s fastest and most efficient model, delivering near-frontier intelligence at a fraction of the cost and latency of larger Claude models.`,context_window:2e5,max_output_tokens:64e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1,output:5,cached_input:.1}},{id:`anthropic/claude-opus-4.1`,name:`Anthropic: Claude Opus 4.1`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-21`,family:`claude-opus`,description:`Claude Opus 4.1 is an updated version of Anthropic’s flagship model, offering improved performance in coding, reasoning, and agentic tasks.`,context_window:2e5,max_output_tokens:32e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:15,output:75,cached_input:1.5}},{id:`anthropic/claude-opus-4.5`,name:`Anthropic: Claude Opus 4.5`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-21`,family:`claude-opus`,description:`Claude Opus 4.5 is Anthropic’s frontier reasoning model optimized for complex software engineering, agentic workflows, and long-horizon computer use.`,context_window:2e5,max_output_tokens:64e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:5,output:25,cached_input:.5}},{id:`anthropic/claude-opus-4.6`,name:`Anthropic: Claude Opus 4.6`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-21`,family:`claude-opus`,description:`Opus 4.6 is Anthropic’s strongest model for coding and long-running professional tasks.`,context_window:1e6,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:5,output:25,cached_input:.5}},{id:`anthropic/claude-opus-4`,name:`Anthropic: Claude Opus 4`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-21`,family:`claude-opus`,description:`Claude Opus 4 is benchmarked as the world’s best coding model, at time of release, bringing sustained performance on complex, long-running tasks and agent workflows.`,context_window:2e5,max_output_tokens:32e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:15,output:75,cached_input:1.5}},{id:`anthropic/claude-sonnet-4.5`,name:`Anthropic: Claude Sonnet 4.5`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-21`,family:`claude-sonnet`,description:`Claude Sonnet 4.5 is Anthropic’s most advanced Sonnet model to date, optimized for real-world agents and coding workflows.`,context_window:1e6,max_output_tokens:64e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:3,output:15,cached_input:.3}},{id:`anthropic/claude-sonnet-4.6`,name:`Anthropic: Claude Sonnet 4.6`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-21`,family:`claude-sonnet`,description:`Sonnet 4.6 is Anthropic's most capable Sonnet-class model yet, with frontier performance across coding, agents, and professional work.`,context_window:1e6,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:3,output:15,cached_input:.3}},{id:`anthropic/claude-sonnet-4`,name:`Anthropic: Claude Sonnet 4`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-21`,family:`claude-sonnet`,description:`Claude Sonnet 4 significantly enhances the capabilities of its predecessor, Sonnet 3.7, excelling in both coding and reasoning tasks with improved precision and controllability.`,context_window:2e5,max_output_tokens:64e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:3,output:15,cached_input:.3}},{id:`arcee-ai/coder-large`,name:`Arcee AI: Coder Large`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-21`,description:`Coder‑Large is a 32 B‑parameter offspring of Qwen 2.5‑Instruct that has been further trained on permissively‑licensed GitHub, CodeSearchNet and synthetic bug‑fix corpora.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.5,output:.8}},{id:`arcee-ai/maestro-reasoning`,name:`Arcee AI: Maestro Reasoning`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-21`,description:`Maestro Reasoning is Arcee's flagship analysis model: a 32 B‑parameter derivative of Qwen 2.5‑32 B tuned with DPO and chain‑of‑thought RL for step‑by‑step logic.`,context_window:131072,max_output_tokens:32e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:.9,output:3.3}},{id:`arcee-ai/spotlight`,name:`Arcee AI: Spotlight`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-21`,description:`Spotlight is a 7‑billion‑parameter vision‑language model derived from Qwen 2.5‑VL and fine‑tuned by Arcee AI for tight image‑text grounding tasks.`,context_window:131072,max_output_tokens:65537,modalities:{input:[`image`,`text`],output:[`text`]},pricing:{input:.18,output:.18}},{id:`arcee-ai/trinity-large-preview:free`,name:`Arcee AI: Trinity Large Preview (free)`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-21`,description:`Trinity-Large-Preview is a frontier-scale open-weight language model from Arcee, built as a 400B-parameter sparse Mixture-of-Experts with 13B active parameters per token using 4-of-256 expert routing.`,context_window:131e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0}},{id:`arcee-ai/trinity-mini:free`,name:`Arcee AI: Trinity Mini (free)`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-21`,description:`Trinity Mini is a 26B-parameter (3B active) sparse mixture-of-experts language model featuring 128 experts with 8 active per token.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0}},{id:`arcee-ai/trinity-mini`,name:`Arcee AI: Trinity Mini`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-21`,description:`Trinity Mini is a 26B-parameter (3B active) sparse mixture-of-experts language model featuring 128 experts with 8 active per token.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.045,output:.15}},{id:`arcee-ai/virtuoso-large`,name:`Arcee AI: Virtuoso Large`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-21`,description:`Virtuoso‑Large is Arcee's top‑tier general‑purpose LLM at 72 B parameters, tuned to tackle cross‑domain reasoning, creative writing and enterprise QA.`,context_window:131072,max_output_tokens:64e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.75,output:1.2}},{id:`baidu/ernie-4.5-21b-a3b-thinking`,name:`Baidu: ERNIE 4.5 21B A3B Thinking`,created_by:`baidu`,source:`official`,last_updated:`2026-03-21`,description:`ERNIE-4.5-21B-A3B-Thinking is Baidu's upgraded lightweight MoE model, refined to boost reasoning depth and quality for top-tier performance in logical puzzles, math, science, coding, text generation, `,context_window:131072,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:.07,output:.28}},{id:`baidu/ernie-4.5-21b-a3b`,name:`Baidu: ERNIE 4.5 21B A3B`,created_by:`baidu`,source:`official`,last_updated:`2026-03-21`,description:`A sophisticated text-based Mixture-of-Experts (MoE) model featuring 21B total parameters with 3B activated per token, delivering exceptional multimodal understanding and generation through heterogeneous MoE structures and modality-isolated routing.`,context_window:12e4,max_output_tokens:8e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.07,output:.28}},{id:`baidu/ernie-4.5-300b-a47b`,name:`Baidu: ERNIE 4.5 300B A47B `,created_by:`baidu`,source:`official`,last_updated:`2026-03-21`,description:`ERNIE-4.5-300B-A47B is a 300B parameter Mixture-of-Experts (MoE) language model developed by Baidu as part of the ERNIE 4.5 series.`,context_window:123e3,max_output_tokens:12e3,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.28,output:1.1}},{id:`baidu/ernie-4.5-vl-28b-a3b`,name:`Baidu: ERNIE 4.5 VL 28B A3B`,created_by:`baidu`,source:`official`,last_updated:`2026-03-21`,description:`A powerful multimodal Mixture-of-Experts chat model featuring 28B total parameters with 3B activated per token, delivering exceptional text and vision understanding through its innovative heterogeneous MoE structure with modality-isolated routing.`,context_window:3e4,max_output_tokens:8e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:.14,output:.56}},{id:`baidu/ernie-4.5-vl-424b-a47b`,name:`Baidu: ERNIE 4.5 VL 424B A47B `,created_by:`baidu`,source:`official`,last_updated:`2026-03-21`,description:`ERNIE-4.5-VL-424B-A47B is a multimodal Mixture-of-Experts (MoE) model from Baidu’s ERNIE 4.5 series, featuring 424B total parameters with 47B active per token.`,context_window:123e3,max_output_tokens:16e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:.42,output:1.25}},{id:`bytedance-seed/seed-1.6-flash`,name:`ByteDance Seed: Seed 1.6 Flash`,created_by:`bytedance-seed`,source:`official`,last_updated:`2026-03-21`,description:`Seed 1.6 Flash is an ultra-fast multimodal deep thinking model by ByteDance Seed, supporting both text and visual understanding.`,context_window:262144,max_output_tokens:32768,modalities:{input:[`image`,`text`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.075,output:.3}},{id:`bytedance-seed/seed-1.6`,name:`ByteDance Seed: Seed 1.6`,created_by:`bytedance-seed`,source:`official`,last_updated:`2026-03-21`,description:`Seed 1.6 is a general-purpose model released by the ByteDance Seed team.`,context_window:262144,max_output_tokens:32768,modalities:{input:[`image`,`text`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.25,output:2}},{id:`bytedance-seed/seed-2.0-lite`,name:`ByteDance Seed: Seed-2.0-Lite`,created_by:`bytedance-seed`,source:`official`,last_updated:`2026-03-21`,description:`Seed-2.0-Lite is a versatile, cost‑efficient enterprise workhorse that delivers strong multimodal and agent capabilities while offering noticeably lower latency, making it a practical default choice for most production workloads across text, vision, and tools.`,context_window:262144,max_output_tokens:131072,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.25,output:2}},{id:`bytedance-seed/seed-2.0-mini`,name:`ByteDance Seed: Seed-2.0-Mini`,created_by:`bytedance-seed`,source:`official`,last_updated:`2026-03-21`,description:`Seed-2.0-mini targets latency-sensitive, high-concurrency, and cost-sensitive scenarios, emphasizing fast response and flexible inference deployment.`,context_window:262144,max_output_tokens:131072,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.1,output:.4}},{id:`bytedance/ui-tars-1.5-7b`,name:`ByteDance: UI-TARS 7B `,created_by:`bytedance`,source:`official`,last_updated:`2026-03-21`,description:`UI-TARS-1.5 is a multimodal vision-language agent optimized for GUI-based environments, including desktop interfaces, web browsers, mobile systems, and games.`,context_window:128e3,max_output_tokens:2048,modalities:{input:[`image`,`text`],output:[`text`]},pricing:{input:.1,output:.2}},{id:`cognitivecomputations/dolphin-mistral-24b-venice-edition:free`,name:`Venice: Uncensored (free)`,created_by:`cognitivecomputations`,source:`official`,last_updated:`2026-03-21`,description:`Venice Uncensored Dolphin Mistral 24B Venice Edition is a fine-tuned variant of Mistral-Small-24B-Instruct-2501, developed by dphn.ai in collaboration with Venice.ai.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0}},{id:`cohere/command-a`,name:`Cohere: Command A`,created_by:`cohere`,source:`official`,last_updated:`2026-03-21`,description:`Command A is an open-weights 111B parameter model with a 256k context window focused on delivering great performance across agentic, multilingual, and coding use cases.`,context_window:256e3,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:2.5,output:10}},{id:`cohere/command-r-08-2024`,name:`Cohere: Command R (08-2024)`,created_by:`cohere`,source:`official`,last_updated:`2026-03-21`,description:`command-r-08-2024 is an update of the [Command R](/models/cohere/command-r) with improved performance for multilingual retrieval-augmented generation (RAG) and tool use.`,context_window:128e3,max_output_tokens:4e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.15,output:.6}},{id:`cohere/command-r-plus-08-2024`,name:`Cohere: Command R+ (08-2024)`,created_by:`cohere`,source:`official`,last_updated:`2026-03-21`,description:`command-r-plus-08-2024 is an update of the [Command R+](/models/cohere/command-r-plus) with roughly 50% higher throughput and 25% lower latencies as compared to the previous Command R+ version, while keeping the hardware footprint the same.`,context_window:128e3,max_output_tokens:4e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2.5,output:10}},{id:`cohere/command-r7b-12-2024`,name:`Cohere: Command R7B (12-2024)`,created_by:`cohere`,source:`official`,last_updated:`2026-03-21`,description:`Command R7B (12-2024) is a small, fast update of the Command R+ model, delivered in December 2024.`,context_window:128e3,max_output_tokens:4e3,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.038,output:.15}},{id:`deepcogito/cogito-v2.1-671b`,name:`Deep Cogito: Cogito v2.1 671B`,created_by:`deepcogito`,source:`official`,last_updated:`2026-03-21`,description:`Cogito v2.1 671B MoE represents one of the strongest open models globally, matching performance of frontier closed and open models.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:1.25,output:1.25}},{id:`deepseek/deepseek-chat-v3-0324`,name:`DeepSeek: DeepSeek V3 0324`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-21`,family:`deepseek-chat`,description:`DeepSeek V3, a 685B-parameter, mixture-of-experts model, is the latest iteration of the flagship chat model family from the DeepSeek team.`,context_window:163840,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.2,output:.77,cached_input:.135}},{id:`deepseek/deepseek-chat-v3.1`,name:`DeepSeek: DeepSeek V3.1`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-21`,family:`deepseek-chat`,description:`DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes via prompt templates.`,context_window:32768,max_output_tokens:7168,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.15,output:.75}},{id:`deepseek/deepseek-chat`,name:`DeepSeek: DeepSeek V3`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-21`,family:`deepseek-chat`,description:`DeepSeek-V3 is the latest model from the DeepSeek team, building upon the instruction following and coding abilities of the previous versions.`,context_window:163840,max_output_tokens:163840,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.32,output:.89}},{id:`deepseek/deepseek-r1-0528`,name:`DeepSeek: R1 0528`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-21`,family:`deepseek-r1`,description:`May 28th update to the [original DeepSeek R1](/deepseek/deepseek-r1) Performance on par with [OpenAI o1](/openai/o1), but open-sourced and with fully open reasoning tokens.`,context_window:163840,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.45,output:2.15,cached_input:.225}},{id:`deepseek/deepseek-r1-distill-llama-70b`,name:`DeepSeek: R1 Distill Llama 70B`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-21`,family:`deepseek-r1`,description:`DeepSeek R1 Distill Llama 70B is a distilled large language model based on [Llama-3.3-70B-Instruct](/meta-llama/llama-3.3-70b-instruct), using outputs from [DeepSeek R1](/deepseek/deepseek-r1).`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.7,output:.8}},{id:`deepseek/deepseek-r1-distill-qwen-32b`,name:`DeepSeek: R1 Distill Qwen 32B`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-21`,family:`deepseek-r1`,description:`DeepSeek R1 Distill Qwen 32B is a distilled large language model based on [Qwen 2.5 32B](https://huggingface.co/Qwen/Qwen2.5-32B), using outputs from [DeepSeek R1](/deepseek/deepseek-r1).`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.29,output:.29}},{id:`deepseek/deepseek-r1`,name:`DeepSeek: R1`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-21`,family:`deepseek-r1`,description:`DeepSeek R1 is here: Performance on par with [OpenAI o1](/openai/o1), but open-sourced and with fully open reasoning tokens.`,context_window:64e3,max_output_tokens:16e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:.7,output:2.5}},{id:`deepseek/deepseek-v3.1-terminus`,name:`DeepSeek: DeepSeek V3.1 Terminus`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-21`,family:`deepseek`,description:`DeepSeek-V3.1 Terminus is an update to [DeepSeek V3.1](/deepseek/deepseek-chat-v3.1) that maintains the model's original capabilities while addressing issues reported by users, including language consistency and agent capabilities, further optimizing the model's performance in coding and search agents.`,context_window:163840,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.21,output:.78,cached_input:.105}},{id:`deepseek/deepseek-v3.2-exp`,name:`DeepSeek: DeepSeek V3.2 Exp`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-21`,family:`deepseek`,description:`DeepSeek-V3.2-Exp is an experimental large language model released by DeepSeek as an intermediate step between V3.1 and future architectures.`,context_window:163840,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.27,output:.41}},{id:`deepseek/deepseek-v3.2-speciale`,name:`DeepSeek: DeepSeek V3.2 Speciale`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-21`,family:`deepseek`,description:`DeepSeek-V3.2-Speciale is a high-compute variant of DeepSeek-V3.2 optimized for maximum reasoning and agentic performance.`,context_window:163840,max_output_tokens:163840,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.4,output:1.2,cached_input:.2}},{id:`deepseek/deepseek-v3.2`,name:`DeepSeek: DeepSeek V3.2`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-21`,family:`deepseek`,description:`DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance.`,context_window:163840,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.26,output:.38,cached_input:.13}},{id:`eleutherai/llemma_7b`,name:`EleutherAI: Llemma 7b`,created_by:`eleutherai`,source:`official`,last_updated:`2026-03-21`,description:`Llemma 7B is a language model for mathematics.`,context_window:4096,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},pricing:{input:.8,output:1.2}},{id:`essentialai/rnj-1-instruct`,name:`EssentialAI: Rnj 1 Instruct`,created_by:`essentialai`,source:`official`,last_updated:`2026-03-21`,description:`Rnj-1 is an 8B-parameter, dense, open-weight model family developed by Essential AI and trained from scratch with a focus on programming, math, and scientific reasoning.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.15,output:.15}},{id:`google/gemini-2.0-flash-001`,name:`Google: Gemini 2.0 Flash`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-2.0`,description:`Gemini Flash 2.0 offers a significantly faster time to first token (TTFT) compared to [Gemini Flash 1.5](/google/gemini-flash-1.5), while maintaining quality on par with larger models like [Gemini Pro 1.5](/google/gemini-pro-1.5).`,status:`deprecated`,deprecation_date:`2026-06-01`,context_window:1048576,max_output_tokens:8192,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.1,output:.4,cached_input:.025}},{id:`google/gemini-2.0-flash-lite-001`,name:`Google: Gemini 2.0 Flash Lite`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-2.0`,description:`Gemini 2.0 Flash Lite offers a significantly faster time to first token (TTFT) compared to [Gemini Flash 1.5](/google/gemini-flash-1.5), while maintaining quality on par with larger models like [Gemin`,status:`deprecated`,deprecation_date:`2026-06-01`,context_window:1048576,max_output_tokens:8192,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.075,output:.3}},{id:`google/gemini-2.5-flash-image`,name:`Google: Nano Banana (Gemini 2.5 Flash Image)`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-2.5`,description:`Gemini 2.5 Flash Image, a.k.a.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`image`,`text`]},capabilities:{structured_output:!0},pricing:{input:.3,output:2.5,cached_input:.03}},{id:`google/gemini-2.5-flash-lite-preview-09-2025`,name:`Google: Gemini 2.5 Flash Lite Preview 09-2025`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-2.5`,description:`Gemini 2.5 Flash-Lite is a lightweight reasoning model in the Gemini 2.5 family, optimized for ultra-low latency and cost efficiency.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.1,output:.4,cached_input:.01}},{id:`google/gemini-2.5-flash-lite`,name:`Google: Gemini 2.5 Flash Lite`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-2.5`,description:`Gemini 2.5 Flash-Lite is a lightweight reasoning model in the Gemini 2.5 family, optimized for ultra-low latency and cost efficiency.`,context_window:1048576,max_output_tokens:65535,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.1,output:.4,cached_input:.01}},{id:`google/gemini-2.5-flash`,name:`Google: Gemini 2.5 Flash`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-2.5`,description:`Gemini 2.5 Flash is Google's state-of-the-art workhorse model, specifically designed for advanced reasoning, coding, mathematics, and scientific tasks.`,context_window:1048576,max_output_tokens:65535,modalities:{input:[`image`,`text`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.3,output:2.5,cached_input:.03}},{id:`google/gemini-2.5-pro-preview-05-06`,name:`Google: Gemini 2.5 Pro Preview 05-06`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-2.5`,description:`Gemini 2.5 Pro is Google’s state-of-the-art AI model designed for advanced reasoning, coding, mathematics, and scientific tasks.`,context_window:1048576,max_output_tokens:65535,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.25,output:10,cached_input:.125}},{id:`google/gemini-2.5-pro-preview`,name:`Google: Gemini 2.5 Pro Preview 06-05`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-2.5`,description:`Gemini 2.5 Pro is Google’s state-of-the-art AI model designed for advanced reasoning, coding, mathematics, and scientific tasks.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`image`,`text`,`audio`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.25,output:10,cached_input:.125}},{id:`google/gemini-2.5-pro`,name:`Google: Gemini 2.5 Pro`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-2.5`,description:`Gemini 2.5 Pro is Google’s state-of-the-art AI model designed for advanced reasoning, coding, mathematics, and scientific tasks.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.25,output:10,cached_input:.125}},{id:`google/gemini-3-flash-preview`,name:`Google: Gemini 3 Flash Preview`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-3`,description:`Gemini 3 Flash Preview is a high speed, high value thinking model designed for agentic workflows, multi turn chat, and coding assistance.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.5,output:3,cached_input:.05}},{id:`google/gemini-3-pro-image-preview`,name:`Google: Nano Banana Pro (Gemini 3 Pro Image Preview)`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-3`,description:`Nano Banana Pro is Google’s most advanced image-generation and editing model, built on Gemini 3 Pro.`,context_window:65536,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`image`,`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:2,output:12,cached_input:.2}},{id:`google/gemini-3-pro-preview`,name:`Google: Gemini 3 Pro Preview`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-3`,description:`Gemini 3 Pro is Google’s flagship frontier model for high-precision multimodal reasoning, combining strong performance across text, image, video, audio, and code with a 1M-token context window.`,status:`deprecated`,deprecation_date:`2026-03-26`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:2,output:12,cached_input:.2}},{id:`google/gemini-3.1-flash-image-preview`,name:`Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview)`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-3.1`,description:`Gemini 3.1 Flash Image Preview, a.k.a.`,context_window:65536,max_output_tokens:65536,modalities:{input:[`image`,`text`],output:[`image`,`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.5,output:3}},{id:`google/gemini-3.1-flash-lite-preview`,name:`Google: Gemini 3.1 Flash Lite Preview`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-3.1`,description:`Gemini 3.1 Flash Lite Preview is Google's high-efficiency model optimized for high-volume use cases.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`,`audio`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.25,output:1.5,cached_input:.025}},{id:`google/gemini-3.1-pro-preview-customtools`,name:`Google: Gemini 3.1 Pro Preview Custom Tools`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-3.1`,description:`Gemini 3.1 Pro Preview Custom Tools is a variant of Gemini 3.1 Pro that improves tool selection behavior by preventing overuse of a general bash tool when more efficient third-party or user-defined functions are available.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`text`,`audio`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:2,output:12,cached_input:.2}},{id:`google/gemini-3.1-pro-preview`,name:`Google: Gemini 3.1 Pro Preview`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemini-3.1`,description:`Gemini 3.1 Pro Preview is Google’s frontier reasoning model, delivering enhanced software engineering performance, improved agentic reliability, and more efficient token usage across complex workflows.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`audio`,`image`,`text`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:2,output:12,cached_input:.2}},{id:`google/gemma-2-27b-it`,name:`Google: Gemma 2 27B`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemma-2`,description:`Gemma 2 27B by Google is an open model built from the same research and technology used to create the [Gemini models](/models?q=gemini).`,context_window:8192,max_output_tokens:2048,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.65,output:.65}},{id:`google/gemma-2-9b-it`,name:`Google: Gemma 2 9B`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemma-2`,description:`Gemma 2 9B by Google is an advanced, open-source language model that sets a new standard for efficiency and performance in its size class.`,context_window:8192,modalities:{input:[`text`],output:[`text`]},pricing:{input:.03,output:.09}},{id:`google/gemma-3-12b-it:free`,name:`Google: Gemma 3 12B (free)`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemma-3`,description:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`,context_window:32768,max_output_tokens:8192,modalities:{input:[`text`,`image`],output:[`text`]}},{id:`google/gemma-3-12b-it`,name:`Google: Gemma 3 12B`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemma-3`,description:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`,context_window:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.04,output:.13}},{id:`google/gemma-3-27b-it:free`,name:`Google: Gemma 3 27B (free)`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemma-3`,description:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`,context_window:131072,max_output_tokens:8192,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0}},{id:`google/gemma-3-27b-it`,name:`Google: Gemma 3 27B`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemma-3`,description:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.08,output:.16}},{id:`google/gemma-3-4b-it:free`,name:`Google: Gemma 3 4B (free)`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemma-3`,description:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`,context_window:32768,max_output_tokens:8192,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0}},{id:`google/gemma-3-4b-it`,name:`Google: Gemma 3 4B`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemma-3`,description:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`,context_window:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.04,output:.08}},{id:`google/gemma-3n-e2b-it:free`,name:`Google: Gemma 3n 2B (free)`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemma-3n`,description:`Gemma 3n E2B IT is a multimodal, instruction-tuned model developed by Google DeepMind, designed to operate efficiently at an effective parameter size of 2B while leveraging a 6B architecture.`,context_window:8192,max_output_tokens:2048,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0}},{id:`google/gemma-3n-e4b-it:free`,name:`Google: Gemma 3n 4B (free)`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemma-3n`,description:`Gemma 3n E4B-it is optimized for efficient execution on mobile and low-resource devices, such as phones, laptops, and tablets.`,context_window:8192,max_output_tokens:2048,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0}},{id:`google/gemma-3n-e4b-it`,name:`Google: Gemma 3n 4B`,created_by:`google`,source:`official`,last_updated:`2026-03-21`,family:`gemma-3n`,description:`Gemma 3n E4B-it is optimized for efficient execution on mobile and low-resource devices, such as phones, laptops, and tablets.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.02,output:.04}},{id:`gryphe/mythomax-l2-13b`,name:`MythoMax 13B`,created_by:`gryphe`,source:`official`,last_updated:`2026-03-21`,description:`One of the highest performing and most popular fine-tunes of Llama 2 13B, with rich descriptions and roleplay.`,context_window:4096,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.06,output:.06}},{id:`ibm-granite/granite-4.0-h-micro`,name:`IBM: Granite 4.0 Micro`,created_by:`ibm-granite`,source:`official`,last_updated:`2026-03-21`,description:`Granite-4.0-H-Micro is a 3B parameter from the Granite 4 family of models.`,context_window:131e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:.017,output:.11}},{id:`inception/mercury-2`,name:`Inception: Mercury 2`,created_by:`inception`,source:`official`,last_updated:`2026-03-21`,description:`Mercury 2 is an extremely fast reasoning LLM, and the first reasoning diffusion LLM (dLLM).`,context_window:128e3,max_output_tokens:5e4,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.25,output:.75,cached_input:.025}},{id:`inception/mercury-coder`,name:`Inception: Mercury Coder`,created_by:`inception`,source:`official`,last_updated:`2026-03-21`,description:`Mercury Coder is the first diffusion large language model (dLLM).`,context_window:128e3,max_output_tokens:32e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.25,output:.75,cached_input:.025}},{id:`inception/mercury`,name:`Inception: Mercury`,created_by:`inception`,source:`official`,last_updated:`2026-03-21`,description:`Mercury is the first diffusion large language model (dLLM).`,status:`deprecated`,deprecation_date:`2026-04-15`,context_window:128e3,max_output_tokens:32e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.25,output:.75,cached_input:.025}},{id:`inflection/inflection-3-pi`,name:`Inflection: Inflection 3 Pi`,created_by:`inflection`,source:`official`,last_updated:`2026-03-21`,description:`Inflection 3 Pi powers Inflection's [Pi](https://pi.ai) chatbot, including backstory, emotional intelligence, productivity, and safety.`,context_window:8e3,max_output_tokens:1024,modalities:{input:[`text`],output:[`text`]},pricing:{input:2.5,output:10}},{id:`inflection/inflection-3-productivity`,name:`Inflection: Inflection 3 Productivity`,created_by:`inflection`,source:`official`,last_updated:`2026-03-21`,description:`Inflection 3 Productivity is optimized for following instructions.`,context_window:8e3,max_output_tokens:1024,modalities:{input:[`text`],output:[`text`]},pricing:{input:2.5,output:10}},{id:`kwaipilot/kat-coder-pro`,name:`Kwaipilot: KAT-Coder-Pro V1`,created_by:`kwaipilot`,source:`official`,last_updated:`2026-03-21`,description:`KAT-Coder-Pro V1 is KwaiKAT's most advanced agentic coding model in the KAT-Coder series.`,context_window:256e3,max_output_tokens:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.207,output:.828,cached_input:.041}},{id:`liquid/lfm-2-24b-a2b`,name:`LiquidAI: LFM2-24B-A2B`,created_by:`liquid`,source:`official`,last_updated:`2026-03-21`,description:`LFM2-24B-A2B is the largest model in the LFM2 family of hybrid architectures designed for efficient on-device deployment.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.03,output:.12}},{id:`liquid/lfm-2.2-6b`,name:`LiquidAI: LFM2-2.6B`,created_by:`liquid`,source:`official`,last_updated:`2026-03-21`,description:`LFM2 is a new generation of hybrid models developed by Liquid AI, specifically designed for edge AI and on-device deployment.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.01,output:.02}},{id:`liquid/lfm-2.5-1.2b-instruct:free`,name:`LiquidAI: LFM2.5-1.2B-Instruct (free)`,created_by:`liquid`,source:`official`,last_updated:`2026-03-21`,description:`LFM2.5-1.2B-Instruct is a compact, high-performance instruction-tuned model built for fast on-device AI.`,context_window:32768,modalities:{input:[`text`],output:[`text`]}},{id:`liquid/lfm-2.5-1.2b-thinking:free`,name:`LiquidAI: LFM2.5-1.2B-Thinking (free)`,created_by:`liquid`,source:`official`,last_updated:`2026-03-21`,description:`LFM2.5-1.2B-Thinking is a lightweight reasoning-focused model optimized for agentic tasks, data extraction, and RAG—while still running comfortably on edge devices.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0}},{id:`liquid/lfm2-8b-a1b`,name:`LiquidAI: LFM2-8B-A1B`,created_by:`liquid`,source:`official`,last_updated:`2026-03-21`,description:`LFM2-8B-A1B is an efficient on-device Mixture-of-Experts (MoE) model from Liquid AI’s LFM2 family, built for fast, high-quality inference on edge hardware.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.01,output:.02}},{id:`mancer/weaver`,name:`Mancer: Weaver (alpha)`,created_by:`mancer`,source:`official`,last_updated:`2026-03-21`,description:`An attempt to recreate Claude-style verbosity, but don't expect the same level of coherence or memory.`,context_window:8e3,max_output_tokens:2e3,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.75,output:1}},{id:`meituan/longcat-flash-chat`,name:`Meituan: LongCat Flash Chat`,created_by:`meituan`,source:`official`,last_updated:`2026-03-21`,description:`LongCat-Flash-Chat is a large-scale Mixture-of-Experts (MoE) model with 560B total parameters, of which 18.6B–31.3B (≈27B on average) are dynamically activated per input.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.2,output:.8,cached_input:.2}},{id:`meta-llama/llama-3-70b-instruct`,name:`Meta: Llama 3 70B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-3`,description:`Meta's latest class of model (Llama 3) launched with a variety of sizes & flavors.`,context_window:8192,max_output_tokens:8e3,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.51,output:.74}},{id:`meta-llama/llama-3-8b-instruct`,name:`Meta: Llama 3 8B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-3`,description:`Meta's latest class of model (Llama 3) launched with a variety of sizes & flavors.`,context_window:8192,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.03,output:.04}},{id:`meta-llama/llama-3.1-405b`,name:`Meta: Llama 3.1 405B (base)`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-3.1`,description:`Meta's latest class of model (Llama 3.1) launched with a variety of sizes & flavors.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:4,output:4}},{id:`meta-llama/llama-3.1-70b-instruct`,name:`Meta: Llama 3.1 70B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-3.1`,description:`Meta's latest class of model (Llama 3.1) launched with a variety of sizes & flavors.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:.4}},{id:`meta-llama/llama-3.1-8b-instruct`,name:`Meta: Llama 3.1 8B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-3.1`,description:`Meta's latest class of model (Llama 3.1) launched with a variety of sizes & flavors.`,context_window:16384,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.02,output:.05}},{id:`meta-llama/llama-3.2-11b-vision-instruct`,name:`Meta: Llama 3.2 11B Vision Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-3.2`,description:`Llama 3.2 11B Vision is a multimodal model with 11 billion parameters, designed to handle tasks combining visual and textual data.`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.049,output:.049}},{id:`meta-llama/llama-3.2-1b-instruct`,name:`Meta: Llama 3.2 1B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-3.2`,description:`Llama 3.2 1B is a 1-billion-parameter language model focused on efficiently performing natural language tasks, such as summarization, dialogue, and multilingual text analysis.`,context_window:6e4,modalities:{input:[`text`],output:[`text`]},pricing:{input:.027,output:.2}},{id:`meta-llama/llama-3.2-3b-instruct:free`,name:`Meta: Llama 3.2 3B Instruct (free)`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-3.2`,description:`Llama 3.2 3B is a 3-billion-parameter multilingual large language model, optimized for advanced natural language processing tasks like dialogue generation, reasoning, and summarization.`,context_window:131072,modalities:{input:[`text`],output:[`text`]}},{id:`meta-llama/llama-3.2-3b-instruct`,name:`Meta: Llama 3.2 3B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-3.2`,description:`Llama 3.2 3B is a 3-billion-parameter multilingual large language model, optimized for advanced natural language processing tasks like dialogue generation, reasoning, and summarization.`,context_window:8e4,modalities:{input:[`text`],output:[`text`]},pricing:{input:.051,output:.34}},{id:`meta-llama/llama-3.3-70b-instruct:free`,name:`Meta: Llama 3.3 70B Instruct (free)`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-3.3`,description:`The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out).`,context_window:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0}},{id:`meta-llama/llama-3.3-70b-instruct`,name:`Meta: Llama 3.3 70B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-3.3`,description:`The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out).`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.1,output:.32}},{id:`meta-llama/llama-4-maverick`,name:`Meta: Llama 4 Maverick`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-4`,description:`Llama 4 Maverick 17B Instruct (128E) is a high-capacity multimodal language model from Meta, built on a mixture-of-experts (MoE) architecture with 128 experts and 17 billion active parameters per forward pass (400B total).`,context_window:1048576,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.15,output:.6}},{id:`meta-llama/llama-4-scout`,name:`Meta: Llama 4 Scout`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-4`,description:`Llama 4 Scout 17B Instruct (16E) is a mixture-of-experts (MoE) language model developed by Meta, activating 17 billion parameters out of a total of 109B.`,context_window:327680,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.08,output:.3}},{id:`meta-llama/llama-guard-3-8b`,name:`Llama Guard 3 8B`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-guard`,description:`Llama Guard 3 is a Llama-3.1-8B pretrained model, fine-tuned for content safety classification.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},pricing:{input:.02,output:.06}},{id:`meta-llama/llama-guard-4-12b`,name:`Meta: Llama Guard 4 12B`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-21`,family:`llama-guard`,description:`Llama Guard 4 is a Llama 4 Scout-derived multimodal pretrained model, fine-tuned for content safety classification.`,context_window:163840,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.18,output:.18}},{id:`microsoft/phi-4`,name:`Microsoft: Phi 4`,created_by:`microsoft`,source:`official`,last_updated:`2026-03-21`,description:`[Microsoft Research](/microsoft) Phi-4 is designed to perform well in complex reasoning tasks and can operate efficiently in situations with limited memory or where quick responses are needed.`,context_window:16384,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.065,output:.14}},{id:`microsoft/wizardlm-2-8x22b`,name:`WizardLM-2 8x22B`,created_by:`microsoft`,source:`official`,last_updated:`2026-03-21`,description:`WizardLM-2 8x22B is Microsoft AI's most advanced Wizard model.`,context_window:65535,max_output_tokens:8e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:.62,output:.62}},{id:`minimax/minimax-01`,name:`MiniMax: MiniMax-01`,created_by:`minimax`,source:`official`,last_updated:`2026-03-21`,description:`MiniMax-01 is a combines MiniMax-Text-01 for text generation and MiniMax-VL-01 for image understanding.`,context_window:1000192,max_output_tokens:1000192,modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:.2,output:1.1}},{id:`minimax/minimax-m1`,name:`MiniMax: MiniMax M1`,created_by:`minimax`,source:`official`,last_updated:`2026-03-21`,description:`MiniMax-M1 is a large-scale, open-weight reasoning model designed for extended context and high-efficiency inference.`,context_window:1e6,max_output_tokens:4e4,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:.4,output:2.2}},{id:`minimax/minimax-m2-her`,name:`MiniMax: MiniMax M2-her`,created_by:`minimax`,source:`official`,last_updated:`2026-03-21`,description:`MiniMax M2-her is a dialogue-first large language model built for immersive roleplay, character-driven chat, and expressive multi-turn conversations.`,context_window:65536,max_output_tokens:2048,modalities:{input:[`text`],output:[`text`]},pricing:{input:.3,output:1.2,cached_input:.03}},{id:`minimax/minimax-m2.1`,name:`MiniMax: MiniMax M2.1`,created_by:`minimax`,source:`official`,last_updated:`2026-03-21`,description:`MiniMax-M2.1 is a lightweight, state-of-the-art large language model optimized for coding, agentic workflows, and modern application development.`,context_window:196608,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.27,output:.95,cached_input:.029}},{id:`minimax/minimax-m2.5:free`,name:`MiniMax: MiniMax M2.5 (free)`,created_by:`minimax`,source:`official`,last_updated:`2026-03-21`,description:`MiniMax-M2.5 is a SOTA large language model designed for real-world productivity.`,context_window:196608,max_output_tokens:196608,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0}},{id:`minimax/minimax-m2.5`,name:`MiniMax: MiniMax M2.5`,created_by:`minimax`,source:`official`,last_updated:`2026-03-21`,description:`MiniMax-M2.5 is a SOTA large language model designed for real-world productivity.`,context_window:196608,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.2,output:1.17,cached_input:.1}},{id:`minimax/minimax-m2.7`,name:`MiniMax: MiniMax M2.7`,created_by:`minimax`,source:`official`,last_updated:`2026-03-21`,description:`MiniMax-M2.7 is a next-generation large language model designed for autonomous, real-world productivity and continuous improvement.`,context_window:204800,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.3,output:1.2,cached_input:.06}},{id:`minimax/minimax-m2`,name:`MiniMax: MiniMax M2`,created_by:`minimax`,source:`official`,last_updated:`2026-03-21`,description:`MiniMax-M2 is a compact, high-efficiency large language model optimized for end-to-end coding and agentic workflows.`,context_window:196608,max_output_tokens:196608,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.255,output:1,cached_input:.03}},{id:`mistralai/codestral-2508`,name:`Mistral: Codestral 2508`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`codestral`,description:`Mistral's cutting-edge language model for coding released end of July 2025.`,context_window:256e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.3,output:.9,cached_input:.03}},{id:`mistralai/devstral-2512`,name:`Mistral: Devstral 2 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`devstral`,description:`Devstral 2 is a state-of-the-art open-source model by Mistral AI specializing in agentic coding.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:2,cached_input:.04}},{id:`mistralai/devstral-medium`,name:`Mistral: Devstral Medium`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`devstral`,description:`Devstral Medium is a high-performance code generation and agentic reasoning model developed jointly by Mistral AI and All Hands AI.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:2,cached_input:.04}},{id:`mistralai/devstral-small`,name:`Mistral: Devstral Small 1.1`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`devstral`,description:`Devstral Small 1.1 is a 24B parameter open-weight language model for software engineering agents, developed by Mistral AI in collaboration with All Hands AI.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.1,output:.3,cached_input:.01}},{id:`mistralai/ministral-14b-2512`,name:`Mistral: Ministral 3 14B 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`ministral`,description:`The largest model in the Ministral 3 family, Ministral 3 14B offers frontier capabilities and performance comparable to its larger Mistral Small 3.2 24B counterpart.`,context_window:262144,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.2,output:.2,cached_input:.02}},{id:`mistralai/ministral-3b-2512`,name:`Mistral: Ministral 3 3B 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`ministral`,description:`The smallest model in the Ministral 3 family, Ministral 3 3B is a powerful, efficient tiny language model with vision capabilities.`,context_window:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.1,output:.1,cached_input:.01}},{id:`mistralai/ministral-8b-2512`,name:`Mistral: Ministral 3 8B 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`ministral`,description:`A balanced model in the Ministral 3 family, Ministral 3 8B is a powerful, efficient tiny language model with vision capabilities.`,context_window:262144,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.15,output:.15,cached_input:.015}},{id:`mistralai/mistral-7b-instruct-v0.1`,name:`Mistral: Mistral 7B Instruct v0.1`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,description:`A 7.3B parameter model that outperforms Llama 2 13B on all benchmarks, with optimizations for speed and context length.`,context_window:2824,modalities:{input:[`text`],output:[`text`]},pricing:{input:.11,output:.19}},{id:`mistralai/mistral-large-2407`,name:`Mistral Large 2407`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mistral-large`,description:`This is Mistral AI's flagship model, Mistral Large 2 (version mistral-large-2407).`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2,output:6,cached_input:.2}},{id:`mistralai/mistral-large-2411`,name:`Mistral Large 2411`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mistral-large`,description:`Mistral Large 2 2411 is an update of [Mistral Large 2](/mistralai/mistral-large) released together with [Pixtral Large 2411](/mistralai/pixtral-large-2411)
+See the launch announcement and benchmark results [here](https:/`,context_window:2e5,max_output_tokens:4096,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,streaming:!0,vision:!0},pricing:{input:.25,output:1.25,cached_input:.03},model_type:`chat`,release_date:`2024-03-13`,tagline:`Claude 3 Haiku is Anthropic's fastest and most compact model for
+near-instant responsiveness. Quick and accurate targeted performance.
+
+See the launch announcement and benchmark results [here](https:/`,tools:[`function_calling`],license:`proprietary`,open_weight:!1},{id:`anthropic/claude-3.5-haiku`,name:`Anthropic: Claude 3.5 Haiku`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-23`,family:`claude-haiku`,description:`Claude 3.5 Haiku features offers enhanced capabilities in speed, coding accuracy, and tool use.`,context_window:2e5,max_output_tokens:8192,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.8,output:4,cached_input:.08},model_type:`chat`,release_date:`2024-11-04`,tagline:`Claude 3.5 Haiku features offers enhanced capabilities in speed, coding accuracy, and tool use.`,tools:[`function_calling`]},{id:`anthropic/claude-3.5-sonnet`,name:`Anthropic: Claude 3.5 Sonnet`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-23`,family:`claude-sonnet`,description:`New Claude 3.5 Sonnet delivers better-than-Opus capabilities, faster-than-Sonnet speeds, at the same Sonnet prices.`,context_window:2e5,max_output_tokens:8192,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:6,output:30,cached_input:.6},model_type:`chat`,release_date:`2024-10-22`,tagline:`New Claude 3.5 Sonnet delivers better-than-Opus capabilities, faster-than-Sonnet speeds, at the same Sonnet prices.`,tools:[`function_calling`]},{id:`anthropic/claude-3.7-sonnet:thinking`,name:`Anthropic: Claude 3.7 Sonnet (thinking)`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-23`,family:`claude-sonnet`,description:`Claude 3.7 Sonnet is an advanced large language model with improved reasoning, coding, and problem-solving capabilities.`,status:`deprecated`,deprecation_date:`2026-05-05`,context_window:2e5,max_output_tokens:64e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:3,output:15,cached_input:.3},model_type:`chat`,release_date:`2025-02-24`,tagline:`Claude 3.7 Sonnet is an advanced large language model with improved reasoning, coding, and problem-solving capabilities.`,tools:[`function_calling`]},{id:`anthropic/claude-3.7-sonnet`,name:`Anthropic: Claude 3.7 Sonnet`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-23`,family:`claude-sonnet`,description:`Claude 3.7 Sonnet is an advanced large language model with improved reasoning, coding, and problem-solving capabilities.`,status:`deprecated`,deprecation_date:`2026-05-05`,context_window:2e5,max_output_tokens:64e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:3,output:15,cached_input:.3},model_type:`chat`,release_date:`2025-02-24`,tagline:`Claude 3.7 Sonnet is an advanced large language model with improved reasoning, coding, and problem-solving capabilities.`,tools:[`function_calling`]},{id:`anthropic/claude-haiku-4.5`,name:`Anthropic: Claude Haiku 4.5`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-23`,family:`claude-haiku`,description:`Claude Haiku 4.5 is Anthropic’s fastest and most efficient model, delivering near-frontier intelligence at a fraction of the cost and latency of larger Claude models.`,context_window:2e5,max_output_tokens:64e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1,output:5,cached_input:.1},model_type:`chat`,release_date:`2025-10-15`,tagline:`Claude Haiku 4.5 is Anthropic’s fastest and most efficient model, delivering near-frontier intelligence at a fraction of the cost and latency of larger Claude models.`,tools:[`function_calling`]},{id:`anthropic/claude-opus-4.1`,name:`Anthropic: Claude Opus 4.1`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-23`,family:`claude-opus`,description:`Claude Opus 4.1 is an updated version of Anthropic’s flagship model, offering improved performance in coding, reasoning, and agentic tasks.`,context_window:2e5,max_output_tokens:32e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:15,output:75,cached_input:1.5},model_type:`chat`,release_date:`2025-08-05`,tagline:`Claude Opus 4.1 is an updated version of Anthropic’s flagship model, offering improved performance in coding, reasoning, and agentic tasks.`,tools:[`function_calling`]},{id:`anthropic/claude-opus-4.5`,name:`Anthropic: Claude Opus 4.5`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-23`,family:`claude-opus`,description:`Claude Opus 4.5 is Anthropic’s frontier reasoning model optimized for complex software engineering, agentic workflows, and long-horizon computer use.`,context_window:2e5,max_output_tokens:64e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:5,output:25,cached_input:.5},model_type:`chat`,release_date:`2025-11-24`,tagline:`Claude Opus 4.5 is Anthropic’s frontier reasoning model optimized for complex software engineering, agentic workflows, and long-horizon computer use.`,tools:[`function_calling`]},{id:`anthropic/claude-opus-4.6`,name:`Anthropic: Claude Opus 4.6`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-23`,family:`claude-opus`,description:`Opus 4.6 is Anthropic’s strongest model for coding and long-running professional tasks.`,context_window:1e6,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:5,output:25,cached_input:.5},model_type:`chat`,release_date:`2026-02-04`,tagline:`Opus 4.6 is Anthropic’s strongest model for coding and long-running professional tasks.`,tools:[`function_calling`]},{id:`anthropic/claude-opus-4`,name:`Anthropic: Claude Opus 4`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-23`,family:`claude-opus`,description:`Claude Opus 4 is benchmarked as the world’s best coding model, at time of release, bringing sustained performance on complex, long-running tasks and agent workflows.`,context_window:2e5,max_output_tokens:32e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:15,output:75,cached_input:1.5},model_type:`chat`,release_date:`2025-05-22`,tagline:`Claude Opus 4 is benchmarked as the world’s best coding model, at time of release, bringing sustained performance on complex, long-running tasks and agent workflows.`,tools:[`function_calling`]},{id:`anthropic/claude-sonnet-4.5`,name:`Anthropic: Claude Sonnet 4.5`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-23`,family:`claude-sonnet`,description:`Claude Sonnet 4.5 is Anthropic’s most advanced Sonnet model to date, optimized for real-world agents and coding workflows.`,context_window:1e6,max_output_tokens:64e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:3,output:15,cached_input:.3},model_type:`chat`,release_date:`2025-09-29`,tagline:`Claude Sonnet 4.5 is Anthropic’s most advanced Sonnet model to date, optimized for real-world agents and coding workflows.`,tools:[`function_calling`]},{id:`anthropic/claude-sonnet-4.6`,name:`Anthropic: Claude Sonnet 4.6`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-23`,family:`claude-sonnet`,description:`Sonnet 4.6 is Anthropic's most capable Sonnet-class model yet, with frontier performance across coding, agents, and professional work.`,context_window:1e6,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:3,output:15,cached_input:.3},model_type:`chat`,release_date:`2026-02-17`,tagline:`Sonnet 4.6 is Anthropic's most capable Sonnet-class model yet, with frontier performance across coding, agents, and professional work.`,tools:[`function_calling`]},{id:`anthropic/claude-sonnet-4`,name:`Anthropic: Claude Sonnet 4`,created_by:`anthropic`,source:`official`,last_updated:`2026-03-23`,family:`claude-sonnet`,description:`Claude Sonnet 4 significantly enhances the capabilities of its predecessor, Sonnet 3.7, excelling in both coding and reasoning tasks with improved precision and controllability.`,context_window:2e5,max_output_tokens:64e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:3,output:15,cached_input:.3},model_type:`chat`,release_date:`2025-05-22`,tagline:`Claude Sonnet 4 significantly enhances the capabilities of its predecessor, Sonnet 3.7, excelling in both coding and reasoning tasks with improved precision and controllability.`,tools:[`function_calling`]},{id:`arcee-ai/coder-large`,name:`Arcee AI: Coder Large`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-23`,description:`Coder‑Large is a 32 B‑parameter offspring of Qwen 2.5‑Instruct that has been further trained on permissively‑licensed GitHub, CodeSearchNet and synthetic bug‑fix corpora.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.5,output:.8},release_date:`2025-05-05`,tagline:`Coder‑Large is a 32 B‑parameter offspring of Qwen 2.5‑Instruct that has been further trained on permissively‑licensed GitHub, CodeSearchNet and synthetic bug‑fix corpora.`},{id:`arcee-ai/maestro-reasoning`,name:`Arcee AI: Maestro Reasoning`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-23`,description:`Maestro Reasoning is Arcee's flagship analysis model: a 32 B‑parameter derivative of Qwen 2.5‑32 B tuned with DPO and chain‑of‑thought RL for step‑by‑step logic.`,context_window:131072,max_output_tokens:32e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:.9,output:3.3},model_type:`reasoning`,release_date:`2025-05-05`,tagline:`Maestro Reasoning is Arcee's flagship analysis model: a 32 B‑parameter derivative of Qwen 2.5‑32 B tuned with DPO and chain‑of‑thought RL for step‑by‑step logic.`},{id:`arcee-ai/spotlight`,name:`Arcee AI: Spotlight`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-23`,description:`Spotlight is a 7‑billion‑parameter vision‑language model derived from Qwen 2.5‑VL and fine‑tuned by Arcee AI for tight image‑text grounding tasks.`,context_window:131072,max_output_tokens:65537,modalities:{input:[`image`,`text`],output:[`text`]},pricing:{input:.18,output:.18},release_date:`2025-05-05`,tagline:`Spotlight is a 7‑billion‑parameter vision‑language model derived from Qwen 2.5‑VL and fine‑tuned by Arcee AI for tight image‑text grounding tasks.`},{id:`arcee-ai/trinity-large-preview:free`,name:`Arcee AI: Trinity Large Preview (free)`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-23`,description:`Trinity-Large-Preview is a frontier-scale open-weight language model from Arcee, built as a 400B-parameter sparse Mixture-of-Experts with 13B active parameters per token using 4-of-256 expert routing.`,context_window:131e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},release_date:`2026-01-27`,tagline:`Trinity-Large-Preview is a frontier-scale open-weight language model from Arcee, built as a 400B-parameter sparse Mixture-of-Experts with 13B active parameters per token using 4-of-256 expert routing.`,tools:[`function_calling`]},{id:`arcee-ai/trinity-mini:free`,name:`Arcee AI: Trinity Mini (free)`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-23`,description:`Trinity Mini is a 26B-parameter (3B active) sparse mixture-of-experts language model featuring 128 experts with 8 active per token.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},release_date:`2025-12-01`,tagline:`Trinity Mini is a 26B-parameter (3B active) sparse mixture-of-experts language model featuring 128 experts with 8 active per token.`,tools:[`function_calling`]},{id:`arcee-ai/trinity-mini`,name:`Arcee AI: Trinity Mini`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-23`,description:`Trinity Mini is a 26B-parameter (3B active) sparse mixture-of-experts language model featuring 128 experts with 8 active per token.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.045,output:.15},release_date:`2025-12-01`,tagline:`Trinity Mini is a 26B-parameter (3B active) sparse mixture-of-experts language model featuring 128 experts with 8 active per token.`,tools:[`function_calling`]},{id:`arcee-ai/virtuoso-large`,name:`Arcee AI: Virtuoso Large`,created_by:`arcee-ai`,source:`official`,last_updated:`2026-03-23`,description:`Virtuoso‑Large is Arcee's top‑tier general‑purpose LLM at 72 B parameters, tuned to tackle cross‑domain reasoning, creative writing and enterprise QA.`,context_window:131072,max_output_tokens:64e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.75,output:1.2},release_date:`2025-05-05`,tagline:`Virtuoso‑Large is Arcee's top‑tier general‑purpose LLM at 72 B parameters, tuned to tackle cross‑domain reasoning, creative writing and enterprise QA.`,tools:[`function_calling`]},{id:`baidu/ernie-4.5-21b-a3b-thinking`,name:`Baidu: ERNIE 4.5 21B A3B Thinking`,created_by:`baidu`,source:`official`,last_updated:`2026-03-23`,description:`ERNIE-4.5-21B-A3B-Thinking is Baidu's upgraded lightweight MoE model, refined to boost reasoning depth and quality for top-tier performance in logical puzzles, math, science, coding, text generation, `,context_window:131072,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:.07,output:.28},parameters:21,active_parameters:3,release_date:`2025-10-09`,tagline:`ERNIE-4.5-21B-A3B-Thinking is Baidu's upgraded lightweight MoE model, refined to boost reasoning depth and quality for top-tier performance in logical puzzles, math, science, coding, text generation, `},{id:`baidu/ernie-4.5-21b-a3b`,name:`Baidu: ERNIE 4.5 21B A3B`,created_by:`baidu`,source:`official`,last_updated:`2026-03-23`,description:`A sophisticated text-based Mixture-of-Experts (MoE) model featuring 21B total parameters with 3B activated per token, delivering exceptional multimodal understanding and generation through heterogeneous MoE structures and modality-isolated routing.`,context_window:12e4,max_output_tokens:8e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.07,output:.28},parameters:21,active_parameters:3,release_date:`2025-08-12`,tagline:`A sophisticated text-based Mixture-of-Experts (MoE) model featuring 21B total parameters with 3B activated per token, delivering exceptional multimodal understanding and generation through heterogeneo`,tools:[`function_calling`]},{id:`baidu/ernie-4.5-300b-a47b`,name:`Baidu: ERNIE 4.5 300B A47B `,created_by:`baidu`,source:`official`,last_updated:`2026-03-23`,description:`ERNIE-4.5-300B-A47B is a 300B parameter Mixture-of-Experts (MoE) language model developed by Baidu as part of the ERNIE 4.5 series.`,context_window:123e3,max_output_tokens:12e3,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.28,output:1.1},parameters:300,active_parameters:47,release_date:`2025-06-30`,tagline:`ERNIE-4.5-300B-A47B is a 300B parameter Mixture-of-Experts (MoE) language model developed by Baidu as part of the ERNIE 4.5 series.`},{id:`baidu/ernie-4.5-vl-28b-a3b`,name:`Baidu: ERNIE 4.5 VL 28B A3B`,created_by:`baidu`,source:`official`,last_updated:`2026-03-23`,description:`A powerful multimodal Mixture-of-Experts chat model featuring 28B total parameters with 3B activated per token, delivering exceptional text and vision understanding through its innovative heterogeneous MoE structure with modality-isolated routing.`,context_window:3e4,max_output_tokens:8e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:.14,output:.56},parameters:28,active_parameters:3,release_date:`2025-08-12`,tagline:`A powerful multimodal Mixture-of-Experts chat model featuring 28B total parameters with 3B activated per token, delivering exceptional text and vision understanding through its innovative heterogeneou`,tools:[`function_calling`]},{id:`baidu/ernie-4.5-vl-424b-a47b`,name:`Baidu: ERNIE 4.5 VL 424B A47B `,created_by:`baidu`,source:`official`,last_updated:`2026-03-23`,description:`ERNIE-4.5-VL-424B-A47B is a multimodal Mixture-of-Experts (MoE) model from Baidu’s ERNIE 4.5 series, featuring 424B total parameters with 47B active per token.`,context_window:123e3,max_output_tokens:16e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:.42,output:1.25},parameters:424,active_parameters:47,release_date:`2025-06-30`,tagline:`ERNIE-4.5-VL-424B-A47B is a multimodal Mixture-of-Experts (MoE) model from Baidu’s ERNIE 4.5 series, featuring 424B total parameters with 47B active per token.`},{id:`bytedance-seed/seed-1.6-flash`,name:`ByteDance Seed: Seed 1.6 Flash`,created_by:`bytedance-seed`,source:`official`,last_updated:`2026-03-23`,description:`Seed 1.6 Flash is an ultra-fast multimodal deep thinking model by ByteDance Seed, supporting both text and visual understanding.`,context_window:262144,max_output_tokens:32768,modalities:{input:[`image`,`text`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.075,output:.3},release_date:`2025-12-23`,tagline:`Seed 1.6 Flash is an ultra-fast multimodal deep thinking model by ByteDance Seed, supporting both text and visual understanding.`,tools:[`function_calling`]},{id:`bytedance-seed/seed-1.6`,name:`ByteDance Seed: Seed 1.6`,created_by:`bytedance-seed`,source:`official`,last_updated:`2026-03-23`,description:`Seed 1.6 is a general-purpose model released by the ByteDance Seed team.`,context_window:262144,max_output_tokens:32768,modalities:{input:[`image`,`text`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.25,output:2},release_date:`2025-12-23`,tagline:`Seed 1.6 is a general-purpose model released by the ByteDance Seed team.`,tools:[`function_calling`]},{id:`bytedance-seed/seed-2.0-lite`,name:`ByteDance Seed: Seed-2.0-Lite`,created_by:`bytedance-seed`,source:`official`,last_updated:`2026-03-23`,description:`Seed-2.0-Lite is a versatile, cost‑efficient enterprise workhorse that delivers strong multimodal and agent capabilities while offering noticeably lower latency, making it a practical default choice for most production workloads across text, vision, and tools.`,context_window:262144,max_output_tokens:131072,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.25,output:2},release_date:`2026-03-10`,tagline:`Seed-2.0-Lite is a versatile, cost‑efficient enterprise workhorse that delivers strong multimodal and agent capabilities while offering noticeably lower latency, making it a practical default choice f`,tools:[`function_calling`]},{id:`bytedance-seed/seed-2.0-mini`,name:`ByteDance Seed: Seed-2.0-Mini`,created_by:`bytedance-seed`,source:`official`,last_updated:`2026-03-23`,description:`Seed-2.0-mini targets latency-sensitive, high-concurrency, and cost-sensitive scenarios, emphasizing fast response and flexible inference deployment.`,context_window:262144,max_output_tokens:131072,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.1,output:.4},release_date:`2026-02-26`,tagline:`Seed-2.0-mini targets latency-sensitive, high-concurrency, and cost-sensitive scenarios, emphasizing fast response and flexible inference deployment.`,tools:[`function_calling`]},{id:`bytedance/ui-tars-1.5-7b`,name:`ByteDance: UI-TARS 7B `,created_by:`bytedance`,source:`official`,last_updated:`2026-03-23`,description:`UI-TARS-1.5 is a multimodal vision-language agent optimized for GUI-based environments, including desktop interfaces, web browsers, mobile systems, and games.`,context_window:128e3,max_output_tokens:2048,modalities:{input:[`image`,`text`],output:[`text`]},pricing:{input:.1,output:.2},parameters:7,release_date:`2025-07-22`,tagline:`UI-TARS-1.5 is a multimodal vision-language agent optimized for GUI-based environments, including desktop interfaces, web browsers, mobile systems, and games.`},{id:`cognitivecomputations/dolphin-mistral-24b-venice-edition:free`,name:`Venice: Uncensored (free)`,created_by:`cognitivecomputations`,source:`official`,last_updated:`2026-03-23`,description:`Venice Uncensored Dolphin Mistral 24B Venice Edition is a fine-tuned variant of Mistral-Small-24B-Instruct-2501, developed by dphn.ai in collaboration with Venice.ai.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},parameters:24,release_date:`2025-07-09`,tagline:`Venice Uncensored Dolphin Mistral 24B Venice Edition is a fine-tuned variant of Mistral-Small-24B-Instruct-2501, developed by dphn.ai in collaboration with Venice.ai.`},{id:`cohere/command-a`,name:`Cohere: Command A`,created_by:`cohere`,source:`official`,last_updated:`2026-03-23`,description:`Command A is an open-weights 111B parameter model with a 256k context window focused on delivering great performance across agentic, multilingual, and coding use cases.`,context_window:256e3,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:2.5,output:10},family:`command-a`,model_type:`chat`,release_date:`2025-03-13`,tagline:`Command A is an open-weights 111B parameter model with a 256k context window focused on delivering great performance across agentic, multilingual, and coding use cases.`},{id:`cohere/command-r-08-2024`,name:`Cohere: Command R (08-2024)`,created_by:`cohere`,source:`official`,last_updated:`2026-03-23`,description:`command-r-08-2024 is an update of the [Command R](/models/cohere/command-r) with improved performance for multilingual retrieval-augmented generation (RAG) and tool use.`,context_window:128e3,max_output_tokens:4e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:.15,output:.6},family:`command-r`,model_type:`chat`,release_date:`2024-08-30`,tagline:`command-r-08-2024 is an update of the [Command R](/models/cohere/command-r) with improved performance for multilingual retrieval-augmented generation (RAG) and tool use.`,tools:[`function_calling`],license:`cc-by-nc-4.0`,open_weight:!0},{id:`cohere/command-r-plus-08-2024`,name:`Cohere: Command R+ (08-2024)`,created_by:`cohere`,source:`official`,last_updated:`2026-03-23`,description:`command-r-plus-08-2024 is an update of the [Command R+](/models/cohere/command-r-plus) with roughly 50% higher throughput and 25% lower latencies as compared to the previous Command R+ version, while keeping the hardware footprint the same.`,context_window:128e3,max_output_tokens:4e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:2.5,output:10},family:`command-r-plus`,model_type:`chat`,release_date:`2024-08-30`,tagline:`command-r-plus-08-2024 is an update of the [Command R+](/models/cohere/command-r-plus) with roughly 50% higher throughput and 25% lower latencies as compared to the previous Command R+ version, while `,tools:[`function_calling`],license:`cc-by-nc-4.0`,open_weight:!0},{id:`cohere/command-r7b-12-2024`,name:`Cohere: Command R7B (12-2024)`,created_by:`cohere`,source:`official`,last_updated:`2026-03-23`,description:`Command R7B (12-2024) is a small, fast update of the Command R+ model, delivered in December 2024.`,context_window:128e3,max_output_tokens:4e3,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,streaming:!0,tool_call:!0,reasoning:!0},pricing:{input:.038,output:.15},family:`command-r`,model_type:`chat`,release_date:`2024-12-14`,tagline:`Command R7B (12-2024) is a small, fast update of the Command R+ model, delivered in December 2024.`,license:`cc-by-nc-4.0`,open_weight:!0},{id:`deepcogito/cogito-v2.1-671b`,name:`Deep Cogito: Cogito v2.1 671B`,created_by:`deepcogito`,source:`official`,last_updated:`2026-03-23`,description:`Cogito v2.1 671B MoE represents one of the strongest open models globally, matching performance of frontier closed and open models.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:1.25,output:1.25},parameters:671,release_date:`2025-11-13`,tagline:`Cogito v2.1 671B MoE represents one of the strongest open models globally, matching performance of frontier closed and open models.`},{id:`deepseek/deepseek-chat-v3-0324`,name:`DeepSeek: DeepSeek V3 0324`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-23`,family:`deepseek-chat`,description:`DeepSeek V3, a 685B-parameter, mixture-of-experts model, is the latest iteration of the flagship chat model family from the DeepSeek team.`,context_window:163840,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.2,output:.77,cached_input:.135},release_date:`2025-03-24`,tagline:`DeepSeek V3, a 685B-parameter, mixture-of-experts model, is the latest iteration of the flagship chat model family from the DeepSeek team.`,tools:[`function_calling`]},{id:`deepseek/deepseek-chat-v3.1`,name:`DeepSeek: DeepSeek V3.1`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-23`,family:`deepseek-chat`,description:`DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes via prompt templates.`,context_window:32768,max_output_tokens:7168,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.15,output:.75},release_date:`2025-08-21`,tagline:`DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes via prompt templates.`,tools:[`function_calling`]},{id:`deepseek/deepseek-chat`,name:`DeepSeek: DeepSeek V3`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-23`,family:`deepseek-chat`,description:`DeepSeek-V3 is the latest model from the DeepSeek team, building upon the instruction following and coding abilities of the previous versions.`,context_window:163840,max_output_tokens:163840,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,json_mode:!0},pricing:{input:.32,output:.89},release_date:`2024-12-26`,tagline:`DeepSeek-V3 is the latest model from the DeepSeek team, building upon the instruction following and coding abilities of the previous versions.`,tools:[`function_calling`],license:`mit`,open_weight:!0},{id:`deepseek/deepseek-r1-0528`,name:`DeepSeek: R1 0528`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-23`,family:`deepseek-r1`,description:`May 28th update to the [original DeepSeek R1](/deepseek/deepseek-r1) Performance on par with [OpenAI o1](/openai/o1), but open-sourced and with fully open reasoning tokens.`,context_window:163840,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.45,output:2.15,cached_input:.225},model_type:`reasoning`,release_date:`2025-05-28`,tagline:`May 28th update to the [original DeepSeek R1](/deepseek/deepseek-r1) Performance on par with [OpenAI o1](/openai/o1), but open-sourced and with fully open reasoning tokens.`,tools:[`function_calling`]},{id:`deepseek/deepseek-r1-distill-llama-70b`,name:`DeepSeek: R1 Distill Llama 70B`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-23`,family:`deepseek-r1`,description:`DeepSeek R1 Distill Llama 70B is a distilled large language model based on [Llama-3.3-70B-Instruct](/meta-llama/llama-3.3-70b-instruct), using outputs from [DeepSeek R1](/deepseek/deepseek-r1).`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.7,output:.8},model_type:`reasoning`,parameters:70,release_date:`2025-01-23`,tagline:`DeepSeek R1 Distill Llama 70B is a distilled large language model based on [Llama-3.3-70B-Instruct](/meta-llama/llama-3.3-70b-instruct), using outputs from [DeepSeek R1](/deepseek/deepseek-r1).`},{id:`deepseek/deepseek-r1-distill-qwen-32b`,name:`DeepSeek: R1 Distill Qwen 32B`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-23`,family:`deepseek-r1`,description:`DeepSeek R1 Distill Qwen 32B is a distilled large language model based on [Qwen 2.5 32B](https://huggingface.co/Qwen/Qwen2.5-32B), using outputs from [DeepSeek R1](/deepseek/deepseek-r1).`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.29,output:.29},model_type:`reasoning`,parameters:32,release_date:`2025-01-29`,tagline:`DeepSeek R1 Distill Qwen 32B is a distilled large language model based on [Qwen 2.5 32B](https://huggingface.co/Qwen/Qwen2.5-32B), using outputs from [DeepSeek R1](/deepseek/deepseek-r1).`},{id:`deepseek/deepseek-r1`,name:`DeepSeek: R1`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-23`,family:`deepseek-r1`,description:`DeepSeek R1 is here: Performance on par with [OpenAI o1](/openai/o1), but open-sourced and with fully open reasoning tokens.`,context_window:64e3,max_output_tokens:16e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:.7,output:2.5},model_type:`reasoning`,release_date:`2025-01-20`,tagline:`DeepSeek R1 is here: Performance on par with [OpenAI o1](/openai/o1), but open-sourced and with fully open reasoning tokens.`,tools:[`function_calling`]},{id:`deepseek/deepseek-v3.1-terminus`,name:`DeepSeek: DeepSeek V3.1 Terminus`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-23`,family:`deepseek`,description:`DeepSeek-V3.1 Terminus is an update to [DeepSeek V3.1](/deepseek/deepseek-chat-v3.1) that maintains the model's original capabilities while addressing issues reported by users, including language consistency and agent capabilities, further optimizing the model's performance in coding and search agents.`,context_window:163840,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.21,output:.79,cached_input:.13},release_date:`2025-09-22`,tagline:`DeepSeek-V3.1 Terminus is an update to [DeepSeek V3.1](/deepseek/deepseek-chat-v3.1) that maintains the model's original capabilities while addressing issues reported by users, including language cons`,tools:[`function_calling`]},{id:`deepseek/deepseek-v3.2-exp`,name:`DeepSeek: DeepSeek V3.2 Exp`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-23`,family:`deepseek`,description:`DeepSeek-V3.2-Exp is an experimental large language model released by DeepSeek as an intermediate step between V3.1 and future architectures.`,context_window:163840,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.27,output:.41},release_date:`2025-09-29`,tagline:`DeepSeek-V3.2-Exp is an experimental large language model released by DeepSeek as an intermediate step between V3.1 and future architectures.`,tools:[`function_calling`]},{id:`deepseek/deepseek-v3.2-speciale`,name:`DeepSeek: DeepSeek V3.2 Speciale`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-23`,family:`deepseek`,description:`DeepSeek-V3.2-Speciale is a high-compute variant of DeepSeek-V3.2 optimized for maximum reasoning and agentic performance.`,context_window:163840,max_output_tokens:163840,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.4,output:1.2,cached_input:.2},release_date:`2025-12-01`,tagline:`DeepSeek-V3.2-Speciale is a high-compute variant of DeepSeek-V3.2 optimized for maximum reasoning and agentic performance.`},{id:`deepseek/deepseek-v3.2`,name:`DeepSeek: DeepSeek V3.2`,created_by:`deepseek`,source:`official`,last_updated:`2026-03-23`,family:`deepseek`,description:`DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance.`,context_window:163840,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.26,output:.38,cached_input:.13},release_date:`2025-12-01`,tagline:`DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance.`,tools:[`function_calling`]},{id:`eleutherai/llemma_7b`,name:`EleutherAI: Llemma 7b`,created_by:`eleutherai`,source:`official`,last_updated:`2026-03-23`,description:`Llemma 7B is a language model for mathematics.`,context_window:4096,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},pricing:{input:.8,output:1.2},parameters:7,release_date:`2025-04-14`,tagline:`Llemma 7B is a language model for mathematics.`},{id:`essentialai/rnj-1-instruct`,name:`EssentialAI: Rnj 1 Instruct`,created_by:`essentialai`,source:`official`,last_updated:`2026-03-23`,description:`Rnj-1 is an 8B-parameter, dense, open-weight model family developed by Essential AI and trained from scratch with a focus on programming, math, and scientific reasoning.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.15,output:.15},release_date:`2025-12-07`,tagline:`Rnj-1 is an 8B-parameter, dense, open-weight model family developed by Essential AI and trained from scratch with a focus on programming, math, and scientific reasoning.`,tools:[`function_calling`]},{id:`google/gemini-2.0-flash-001`,name:`Google: Gemini 2.0 Flash`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-2.0`,description:`Gemini Flash 2.0 offers a significantly faster time to first token (TTFT) compared to [Gemini Flash 1.5](/google/gemini-flash-1.5), while maintaining quality on par with larger models like [Gemini Pro 1.5](/google/gemini-pro-1.5).`,status:`deprecated`,deprecation_date:`2026-06-01`,context_window:1048576,max_output_tokens:8192,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:.1,output:.4,cached_input:.025},model_type:`chat`,release_date:`2025-02-05`,tagline:`Gemini Flash 2.0 offers a significantly faster time to first token (TTFT) compared to [Gemini Flash 1.5](/google/gemini-flash-1.5), while maintaining quality on par with larger models like [Gemini Pro`,tools:[`function_calling`],license:`proprietary`,open_weight:!1},{id:`google/gemini-2.0-flash-lite-001`,name:`Google: Gemini 2.0 Flash Lite`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-2.0`,description:`Gemini 2.0 Flash Lite offers a significantly faster time to first token (TTFT) compared to [Gemini Flash 1.5](/google/gemini-flash-1.5), while maintaining quality on par with larger models like [Gemin`,status:`deprecated`,deprecation_date:`2026-06-01`,context_window:1048576,max_output_tokens:8192,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:.075,output:.3},model_type:`chat`,release_date:`2025-02-25`,tagline:`Gemini 2.0 Flash Lite offers a significantly faster time to first token (TTFT) compared to [Gemini Flash 1.5](/google/gemini-flash-1.5), while maintaining quality on par with larger models like [Gemin`,tools:[`function_calling`],license:`proprietary`,open_weight:!1},{id:`google/gemini-2.5-flash-image`,name:`Google: Nano Banana (Gemini 2.5 Flash Image)`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-2.5`,description:`Gemini 2.5 Flash Image, a.k.a.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`image`,`text`]},capabilities:{structured_output:!0,streaming:!0,tool_call:!0,reasoning:!0,batch:!0,vision:!0},pricing:{input:.3,output:2.5,cached_input:.03},model_type:`chat`,release_date:`2025-10-07`,tagline:`Gemini 2.5 Flash Image, a.k.a.`,knowledge_cutoff:`2025-06`,license:`proprietary`,open_weight:!1},{id:`google/gemini-2.5-flash-lite-preview-09-2025`,name:`Google: Gemini 2.5 Flash Lite Preview 09-2025`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-2.5`,description:`Gemini 2.5 Flash-Lite is a lightweight reasoning model in the Gemini 2.5 family, optimized for ultra-low latency and cost efficiency.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,batch:!0,vision:!0},pricing:{input:.1,output:.4,cached_input:.01},model_type:`chat`,release_date:`2025-09-25`,tagline:`Gemini 2.5 Flash-Lite is a lightweight reasoning model in the Gemini 2.5 family, optimized for ultra-low latency and cost efficiency.`,tools:[`function_calling`],knowledge_cutoff:`2025-01`,license:`proprietary`,open_weight:!1},{id:`google/gemini-2.5-flash-lite`,name:`Google: Gemini 2.5 Flash Lite`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-2.5`,description:`Gemini 2.5 Flash-Lite is a lightweight reasoning model in the Gemini 2.5 family, optimized for ultra-low latency and cost efficiency.`,context_window:1048576,max_output_tokens:65535,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,batch:!0,vision:!0},pricing:{input:.1,output:.4,cached_input:.01},model_type:`chat`,release_date:`2025-07-22`,tagline:`Gemini 2.5 Flash-Lite is a lightweight reasoning model in the Gemini 2.5 family, optimized for ultra-low latency and cost efficiency.`,tools:[`function_calling`],knowledge_cutoff:`2025-01`,license:`proprietary`,open_weight:!1},{id:`google/gemini-2.5-flash`,name:`Google: Gemini 2.5 Flash`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-2.5`,description:`Gemini 2.5 Flash is Google's state-of-the-art workhorse model, specifically designed for advanced reasoning, coding, mathematics, and scientific tasks.`,context_window:1048576,max_output_tokens:65535,modalities:{input:[`image`,`text`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,batch:!0,vision:!0},pricing:{input:.3,output:2.5,cached_input:.03},model_type:`chat`,release_date:`2025-06-17`,tagline:`Gemini 2.5 Flash is Google's state-of-the-art workhorse model, specifically designed for advanced reasoning, coding, mathematics, and scientific tasks.`,tools:[`function_calling`],knowledge_cutoff:`2025-01`,license:`proprietary`,open_weight:!1},{id:`google/gemini-2.5-pro-preview-05-06`,name:`Google: Gemini 2.5 Pro Preview 05-06`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-2.5`,description:`Gemini 2.5 Pro is Google’s state-of-the-art AI model designed for advanced reasoning, coding, mathematics, and scientific tasks.`,context_window:1048576,max_output_tokens:65535,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:1.25,output:10,cached_input:.125},model_type:`chat`,release_date:`2025-05-07`,tagline:`Gemini 2.5 Pro is Google’s state-of-the-art AI model designed for advanced reasoning, coding, mathematics, and scientific tasks.`,tools:[`function_calling`],license:`proprietary`,open_weight:!1},{id:`google/gemini-2.5-pro-preview`,name:`Google: Gemini 2.5 Pro Preview 06-05`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-2.5`,description:`Gemini 2.5 Pro is Google’s state-of-the-art AI model designed for advanced reasoning, coding, mathematics, and scientific tasks.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`image`,`text`,`audio`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.25,output:10,cached_input:.125},model_type:`chat`,release_date:`2025-06-05`,tagline:`Gemini 2.5 Pro is Google’s state-of-the-art AI model designed for advanced reasoning, coding, mathematics, and scientific tasks.`,tools:[`function_calling`]},{id:`google/gemini-2.5-pro`,name:`Google: Gemini 2.5 Pro`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-2.5`,description:`Gemini 2.5 Pro is Google’s state-of-the-art AI model designed for advanced reasoning, coding, mathematics, and scientific tasks.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,batch:!0,vision:!0},pricing:{input:1.25,output:10,cached_input:.125},model_type:`chat`,release_date:`2025-06-17`,tagline:`Gemini 2.5 Pro is Google’s state-of-the-art AI model designed for advanced reasoning, coding, mathematics, and scientific tasks.`,tools:[`function_calling`],knowledge_cutoff:`2025-01`,license:`proprietary`,open_weight:!1},{id:`google/gemini-3-flash-preview`,name:`Google: Gemini 3 Flash Preview`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-3`,description:`Gemini 3 Flash Preview is a high speed, high value thinking model designed for agentic workflows, multi turn chat, and coding assistance.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,batch:!0,vision:!0},pricing:{input:.5,output:3,cached_input:.05},model_type:`chat`,release_date:`2025-12-17`,tagline:`Gemini 3 Flash Preview is a high speed, high value thinking model designed for agentic workflows, multi turn chat, and coding assistance.`,tools:[`function_calling`],knowledge_cutoff:`2025-01`,license:`proprietary`,open_weight:!1},{id:`google/gemini-3-pro-image-preview`,name:`Google: Nano Banana Pro (Gemini 3 Pro Image Preview)`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-3`,description:`Nano Banana Pro is Google’s most advanced image-generation and editing model, built on Gemini 3 Pro.`,context_window:65536,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`image`,`text`]},capabilities:{structured_output:!0,reasoning:!0,streaming:!0,tool_call:!0,batch:!0,vision:!0},pricing:{input:2,output:12,cached_input:.2},model_type:`chat`,release_date:`2025-11-20`,tagline:`Nano Banana Pro is Google’s most advanced image-generation and editing model, built on Gemini 3 Pro.`,knowledge_cutoff:`2025-01`,license:`proprietary`,open_weight:!1},{id:`google/gemini-3-pro-preview`,name:`Google: Gemini 3 Pro Preview`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-3`,description:`Gemini 3 Pro is Google’s flagship frontier model for high-precision multimodal reasoning, combining strong performance across text, image, video, audio, and code with a 1M-token context window.`,status:`deprecated`,deprecation_date:`2026-03-26`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,batch:!0,vision:!0},pricing:{input:2,output:12,cached_input:.2},model_type:`chat`,release_date:`2025-11-18`,tagline:`Gemini 3 Pro is Google’s flagship frontier model for high-precision multimodal reasoning, combining strong performance across text, image, video, audio, and code with a 1M-token context window.`,tools:[`function_calling`],knowledge_cutoff:`2025-01`,license:`proprietary`,open_weight:!1},{id:`google/gemini-3.1-flash-image-preview`,name:`Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview)`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-3.1`,description:`Gemini 3.1 Flash Image Preview, a.k.a.`,context_window:65536,max_output_tokens:65536,modalities:{input:[`image`,`text`],output:[`image`,`text`]},capabilities:{structured_output:!0,reasoning:!0,streaming:!0,tool_call:!0,batch:!0,vision:!0},pricing:{input:.5,output:3},model_type:`chat`,release_date:`2026-02-26`,tagline:`Gemini 3.1 Flash Image Preview, a.k.a.`,knowledge_cutoff:`2025-01`,license:`proprietary`,open_weight:!1},{id:`google/gemini-3.1-flash-lite-preview`,name:`Google: Gemini 3.1 Flash Lite Preview`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-3.1`,description:`Gemini 3.1 Flash Lite Preview is Google's high-efficiency model optimized for high-volume use cases.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`,`audio`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,batch:!0,vision:!0},pricing:{input:.25,output:1.5,cached_input:.025},model_type:`chat`,release_date:`2026-03-03`,tagline:`Gemini 3.1 Flash Lite Preview is Google's high-efficiency model optimized for high-volume use cases.`,tools:[`function_calling`],knowledge_cutoff:`2025-01`,license:`proprietary`,open_weight:!1},{id:`google/gemini-3.1-pro-preview-customtools`,name:`Google: Gemini 3.1 Pro Preview Custom Tools`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-3.1`,description:`Gemini 3.1 Pro Preview Custom Tools is a variant of Gemini 3.1 Pro that improves tool selection behavior by preventing overuse of a general bash tool when more efficient third-party or user-defined functions are available.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`text`,`audio`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:2,output:12,cached_input:.2},model_type:`chat`,release_date:`2026-02-25`,tagline:`Gemini 3.1 Pro Preview Custom Tools is a variant of Gemini 3.1 Pro that improves tool selection behavior by preventing overuse of a general bash tool when more efficient third-party or user-defined fu`,tools:[`function_calling`]},{id:`google/gemini-3.1-pro-preview`,name:`Google: Gemini 3.1 Pro Preview`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemini-3.1`,description:`Gemini 3.1 Pro Preview is Google’s frontier reasoning model, delivering enhanced software engineering performance, improved agentic reliability, and more efficient token usage across complex workflows.`,context_window:1048576,max_output_tokens:65536,modalities:{input:[`audio`,`image`,`text`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,batch:!0,vision:!0},pricing:{input:2,output:12,cached_input:.2},model_type:`chat`,release_date:`2026-02-19`,tagline:`Gemini 3.1 Pro Preview is Google’s frontier reasoning model, delivering enhanced software engineering performance, improved agentic reliability, and more efficient token usage across complex workflows`,tools:[`function_calling`],knowledge_cutoff:`2025-01`,license:`proprietary`,open_weight:!1},{id:`google/gemma-2-27b-it`,name:`Google: Gemma 2 27B`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemma-2`,description:`Gemma 2 27B by Google is an open model built from the same research and technology used to create the [Gemini models](/models?q=gemini).`,context_window:8192,max_output_tokens:2048,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.65,output:.65},model_type:`chat`,parameters:27,release_date:`2024-07-13`,tagline:`Gemma 2 27B by Google is an open model built from the same research and technology used to create the [Gemini models](/models?q=gemini).`},{id:`google/gemma-2-9b-it`,name:`Google: Gemma 2 9B`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemma-2`,description:`Gemma 2 9B by Google is an advanced, open-source language model that sets a new standard for efficiency and performance in its size class.`,context_window:8192,modalities:{input:[`text`],output:[`text`]},pricing:{input:.03,output:.09},model_type:`chat`,parameters:9,release_date:`2024-06-28`,tagline:`Gemma 2 9B by Google is an advanced, open-source language model that sets a new standard for efficiency and performance in its size class.`},{id:`google/gemma-3-12b-it:free`,name:`Google: Gemma 3 12B (free)`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemma-3`,description:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`,context_window:32768,max_output_tokens:8192,modalities:{input:[`text`,`image`],output:[`text`]},model_type:`chat`,parameters:12,release_date:`2025-03-13`,tagline:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`},{id:`google/gemma-3-12b-it`,name:`Google: Gemma 3 12B`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemma-3`,description:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`,context_window:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.04,output:.13},model_type:`chat`,parameters:12,release_date:`2025-03-13`,tagline:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`},{id:`google/gemma-3-27b-it:free`,name:`Google: Gemma 3 27B (free)`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemma-3`,description:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`,context_window:131072,max_output_tokens:8192,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},model_type:`chat`,parameters:27,release_date:`2025-03-12`,tagline:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`},{id:`google/gemma-3-27b-it`,name:`Google: Gemma 3 27B`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemma-3`,description:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.08,output:.16},model_type:`chat`,parameters:27,release_date:`2025-03-12`,tagline:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`},{id:`google/gemma-3-4b-it:free`,name:`Google: Gemma 3 4B (free)`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemma-3`,description:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`,context_window:32768,max_output_tokens:8192,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},model_type:`chat`,parameters:4,release_date:`2025-03-13`,tagline:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`},{id:`google/gemma-3-4b-it`,name:`Google: Gemma 3 4B`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemma-3`,description:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`,context_window:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.04,output:.08},model_type:`chat`,parameters:4,release_date:`2025-03-13`,tagline:`Gemma 3 introduces multimodality, supporting vision-language input and text outputs.`},{id:`google/gemma-3n-e2b-it:free`,name:`Google: Gemma 3n 2B (free)`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemma-3n`,description:`Gemma 3n E2B IT is a multimodal, instruction-tuned model developed by Google DeepMind, designed to operate efficiently at an effective parameter size of 2B while leveraging a 6B architecture.`,context_window:8192,max_output_tokens:2048,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},model_type:`chat`,release_date:`2025-07-09`,tagline:`Gemma 3n E2B IT is a multimodal, instruction-tuned model developed by Google DeepMind, designed to operate efficiently at an effective parameter size of 2B while leveraging a 6B architecture.`},{id:`google/gemma-3n-e4b-it:free`,name:`Google: Gemma 3n 4B (free)`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemma-3n`,description:`Gemma 3n E4B-it is optimized for efficient execution on mobile and low-resource devices, such as phones, laptops, and tablets.`,context_window:8192,max_output_tokens:2048,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},model_type:`chat`,release_date:`2025-05-20`,tagline:`Gemma 3n E4B-it is optimized for efficient execution on mobile and low-resource devices, such as phones, laptops, and tablets.`},{id:`google/gemma-3n-e4b-it`,name:`Google: Gemma 3n 4B`,created_by:`google`,source:`official`,last_updated:`2026-03-23`,family:`gemma-3n`,description:`Gemma 3n E4B-it is optimized for efficient execution on mobile and low-resource devices, such as phones, laptops, and tablets.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.02,output:.04},model_type:`chat`,release_date:`2025-05-20`,tagline:`Gemma 3n E4B-it is optimized for efficient execution on mobile and low-resource devices, such as phones, laptops, and tablets.`},{id:`gryphe/mythomax-l2-13b`,name:`MythoMax 13B`,created_by:`gryphe`,source:`official`,last_updated:`2026-03-23`,description:`One of the highest performing and most popular fine-tunes of Llama 2 13B, with rich descriptions and roleplay.`,context_window:4096,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.06,output:.06},parameters:13,release_date:`2023-07-02`,tagline:`One of the highest performing and most popular fine-tunes of Llama 2 13B, with rich descriptions and roleplay.`},{id:`ibm-granite/granite-4.0-h-micro`,name:`IBM: Granite 4.0 Micro`,created_by:`ibm-granite`,source:`official`,last_updated:`2026-03-23`,description:`Granite-4.0-H-Micro is a 3B parameter from the Granite 4 family of models.`,context_window:131e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:.017,output:.11},release_date:`2025-10-20`,tagline:`Granite-4.0-H-Micro is a 3B parameter from the Granite 4 family of models.`},{id:`inception/mercury-2`,name:`Inception: Mercury 2`,created_by:`inception`,source:`official`,last_updated:`2026-03-23`,description:`Mercury 2 is an extremely fast reasoning LLM, and the first reasoning diffusion LLM (dLLM).`,context_window:128e3,max_output_tokens:5e4,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.25,output:.75,cached_input:.025},release_date:`2026-03-04`,tagline:`Mercury 2 is an extremely fast reasoning LLM, and the first reasoning diffusion LLM (dLLM).`,tools:[`function_calling`]},{id:`inception/mercury-coder`,name:`Inception: Mercury Coder`,created_by:`inception`,source:`official`,last_updated:`2026-03-23`,description:`Mercury Coder is the first diffusion large language model (dLLM).`,context_window:128e3,max_output_tokens:32e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.25,output:.75,cached_input:.025},release_date:`2025-04-30`,tagline:`Mercury Coder is the first diffusion large language model (dLLM).`,tools:[`function_calling`]},{id:`inception/mercury`,name:`Inception: Mercury`,created_by:`inception`,source:`official`,last_updated:`2026-03-23`,description:`Mercury is the first diffusion large language model (dLLM).`,status:`deprecated`,deprecation_date:`2026-04-15`,context_window:128e3,max_output_tokens:32e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.25,output:.75,cached_input:.025},release_date:`2025-06-26`,tagline:`Mercury is the first diffusion large language model (dLLM).`,tools:[`function_calling`]},{id:`inflection/inflection-3-pi`,name:`Inflection: Inflection 3 Pi`,created_by:`inflection`,source:`official`,last_updated:`2026-03-23`,description:`Inflection 3 Pi powers Inflection's [Pi](https://pi.ai) chatbot, including backstory, emotional intelligence, productivity, and safety.`,context_window:8e3,max_output_tokens:1024,modalities:{input:[`text`],output:[`text`]},pricing:{input:2.5,output:10},release_date:`2024-10-11`,tagline:`Inflection 3 Pi powers Inflection's [Pi](https://pi.ai) chatbot, including backstory, emotional intelligence, productivity, and safety.`},{id:`inflection/inflection-3-productivity`,name:`Inflection: Inflection 3 Productivity`,created_by:`inflection`,source:`official`,last_updated:`2026-03-23`,description:`Inflection 3 Productivity is optimized for following instructions.`,context_window:8e3,max_output_tokens:1024,modalities:{input:[`text`],output:[`text`]},pricing:{input:2.5,output:10},release_date:`2024-10-11`,tagline:`Inflection 3 Productivity is optimized for following instructions.`},{id:`kwaipilot/kat-coder-pro`,name:`Kwaipilot: KAT-Coder-Pro V1`,created_by:`kwaipilot`,source:`official`,last_updated:`2026-03-23`,description:`KAT-Coder-Pro V1 is KwaiKAT's most advanced agentic coding model in the KAT-Coder series.`,context_window:256e3,max_output_tokens:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.207,output:.828,cached_input:.041},release_date:`2025-11-10`,tagline:`KAT-Coder-Pro V1 is KwaiKAT's most advanced agentic coding model in the KAT-Coder series.`,tools:[`function_calling`]},{id:`liquid/lfm-2-24b-a2b`,name:`LiquidAI: LFM2-24B-A2B`,created_by:`liquid`,source:`official`,last_updated:`2026-03-23`,description:`LFM2-24B-A2B is the largest model in the LFM2 family of hybrid architectures designed for efficient on-device deployment.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.03,output:.12},parameters:24,active_parameters:2,release_date:`2026-02-25`,tagline:`LFM2-24B-A2B is the largest model in the LFM2 family of hybrid architectures designed for efficient on-device deployment.`},{id:`liquid/lfm-2.2-6b`,name:`LiquidAI: LFM2-2.6B`,created_by:`liquid`,source:`official`,last_updated:`2026-03-23`,description:`LFM2 is a new generation of hybrid models developed by Liquid AI, specifically designed for edge AI and on-device deployment.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.01,output:.02},parameters:6,release_date:`2025-10-20`,tagline:`LFM2 is a new generation of hybrid models developed by Liquid AI, specifically designed for edge AI and on-device deployment.`},{id:`liquid/lfm-2.5-1.2b-instruct:free`,name:`LiquidAI: LFM2.5-1.2B-Instruct (free)`,created_by:`liquid`,source:`official`,last_updated:`2026-03-23`,description:`LFM2.5-1.2B-Instruct is a compact, high-performance instruction-tuned model built for fast on-device AI.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},parameters:1.2,release_date:`2026-01-20`,tagline:`LFM2.5-1.2B-Instruct is a compact, high-performance instruction-tuned model built for fast on-device AI.`},{id:`liquid/lfm-2.5-1.2b-thinking:free`,name:`LiquidAI: LFM2.5-1.2B-Thinking (free)`,created_by:`liquid`,source:`official`,last_updated:`2026-03-23`,description:`LFM2.5-1.2B-Thinking is a lightweight reasoning-focused model optimized for agentic tasks, data extraction, and RAG—while still running comfortably on edge devices.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0},parameters:1.2,release_date:`2026-01-20`,tagline:`LFM2.5-1.2B-Thinking is a lightweight reasoning-focused model optimized for agentic tasks, data extraction, and RAG—while still running comfortably on edge devices.`},{id:`liquid/lfm2-8b-a1b`,name:`LiquidAI: LFM2-8B-A1B`,created_by:`liquid`,source:`official`,last_updated:`2026-03-23`,description:`LFM2-8B-A1B is an efficient on-device Mixture-of-Experts (MoE) model from Liquid AI’s LFM2 family, built for fast, high-quality inference on edge hardware.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.01,output:.02},parameters:8,active_parameters:1,release_date:`2025-10-20`,tagline:`LFM2-8B-A1B is an efficient on-device Mixture-of-Experts (MoE) model from Liquid AI’s LFM2 family, built for fast, high-quality inference on edge hardware.`},{id:`mancer/weaver`,name:`Mancer: Weaver (alpha)`,created_by:`mancer`,source:`official`,last_updated:`2026-03-23`,description:`An attempt to recreate Claude-style verbosity, but don't expect the same level of coherence or memory.`,context_window:8e3,max_output_tokens:2e3,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.75,output:1},release_date:`2023-08-02`,tagline:`An attempt to recreate Claude-style verbosity, but don't expect the same level of coherence or memory.`},{id:`meituan/longcat-flash-chat`,name:`Meituan: LongCat Flash Chat`,created_by:`meituan`,source:`official`,last_updated:`2026-03-23`,description:`LongCat-Flash-Chat is a large-scale Mixture-of-Experts (MoE) model with 560B total parameters, of which 18.6B–31.3B (≈27B on average) are dynamically activated per input.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.2,output:.8,cached_input:.2},release_date:`2025-09-09`,tagline:`LongCat-Flash-Chat is a large-scale Mixture-of-Experts (MoE) model with 560B total parameters, of which 18.6B–31.3B (≈27B on average) are dynamically activated per input.`,tools:[`function_calling`]},{id:`meta-llama/llama-3-70b-instruct`,name:`Meta: Llama 3 70B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-3`,description:`Meta's latest class of model (Llama 3) launched with a variety of sizes & flavors.`,context_window:8192,max_output_tokens:8e3,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.51,output:.74},model_type:`chat`,parameters:70,release_date:`2024-04-18`,tagline:`Meta's latest class of model (Llama 3) launched with a variety of sizes & flavors.`},{id:`meta-llama/llama-3-8b-instruct`,name:`Meta: Llama 3 8B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-3`,description:`Meta's latest class of model (Llama 3) launched with a variety of sizes & flavors.`,context_window:8192,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.03,output:.04},model_type:`chat`,parameters:8,release_date:`2024-04-18`,tagline:`Meta's latest class of model (Llama 3) launched with a variety of sizes & flavors.`,tools:[`function_calling`]},{id:`meta-llama/llama-3.1-405b`,name:`Meta: Llama 3.1 405B (base)`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-3.1`,description:`Meta's latest class of model (Llama 3.1) launched with a variety of sizes & flavors.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:4,output:4},model_type:`chat`,parameters:405,release_date:`2024-08-02`,tagline:`Meta's latest class of model (Llama 3.1) launched with a variety of sizes & flavors.`},{id:`meta-llama/llama-3.1-70b-instruct`,name:`Meta: Llama 3.1 70B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-3.1`,description:`Meta's latest class of model (Llama 3.1) launched with a variety of sizes & flavors.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:.4},model_type:`chat`,parameters:70,release_date:`2024-07-23`,tagline:`Meta's latest class of model (Llama 3.1) launched with a variety of sizes & flavors.`,tools:[`function_calling`]},{id:`meta-llama/llama-3.1-8b-instruct`,name:`Meta: Llama 3.1 8B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-3.1`,description:`Meta's latest class of model (Llama 3.1) launched with a variety of sizes & flavors.`,context_window:16384,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.02,output:.05},model_type:`chat`,parameters:8,release_date:`2024-07-23`,tagline:`Meta's latest class of model (Llama 3.1) launched with a variety of sizes & flavors.`,tools:[`function_calling`]},{id:`meta-llama/llama-3.2-11b-vision-instruct`,name:`Meta: Llama 3.2 11B Vision Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-3.2`,description:`Llama 3.2 11B Vision is a multimodal model with 11 billion parameters, designed to handle tasks combining visual and textual data.`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.049,output:.049},model_type:`chat`,parameters:11,release_date:`2024-09-25`,tagline:`Llama 3.2 11B Vision is a multimodal model with 11 billion parameters, designed to handle tasks combining visual and textual data.`},{id:`meta-llama/llama-3.2-1b-instruct`,name:`Meta: Llama 3.2 1B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-3.2`,description:`Llama 3.2 1B is a 1-billion-parameter language model focused on efficiently performing natural language tasks, such as summarization, dialogue, and multilingual text analysis.`,context_window:6e4,modalities:{input:[`text`],output:[`text`]},pricing:{input:.027,output:.2},model_type:`chat`,parameters:1,release_date:`2024-09-25`,tagline:`Llama 3.2 1B is a 1-billion-parameter language model focused on efficiently performing natural language tasks, such as summarization, dialogue, and multilingual text analysis.`},{id:`meta-llama/llama-3.2-3b-instruct:free`,name:`Meta: Llama 3.2 3B Instruct (free)`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-3.2`,description:`Llama 3.2 3B is a 3-billion-parameter multilingual large language model, optimized for advanced natural language processing tasks like dialogue generation, reasoning, and summarization.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},model_type:`chat`,parameters:3,release_date:`2024-09-25`,tagline:`Llama 3.2 3B is a 3-billion-parameter multilingual large language model, optimized for advanced natural language processing tasks like dialogue generation, reasoning, and summarization.`},{id:`meta-llama/llama-3.2-3b-instruct`,name:`Meta: Llama 3.2 3B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-3.2`,description:`Llama 3.2 3B is a 3-billion-parameter multilingual large language model, optimized for advanced natural language processing tasks like dialogue generation, reasoning, and summarization.`,context_window:8e4,modalities:{input:[`text`],output:[`text`]},pricing:{input:.051,output:.34},model_type:`chat`,parameters:3,release_date:`2024-09-25`,tagline:`Llama 3.2 3B is a 3-billion-parameter multilingual large language model, optimized for advanced natural language processing tasks like dialogue generation, reasoning, and summarization.`},{id:`meta-llama/llama-3.3-70b-instruct:free`,name:`Meta: Llama 3.3 70B Instruct (free)`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-3.3`,description:`The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out).`,context_window:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},model_type:`chat`,parameters:70,release_date:`2024-12-06`,tagline:`The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out).`,tools:[`function_calling`]},{id:`meta-llama/llama-3.3-70b-instruct`,name:`Meta: Llama 3.3 70B Instruct`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-3.3`,description:`The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out).`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.1,output:.32},model_type:`chat`,parameters:70,release_date:`2024-12-06`,tagline:`The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out).`,tools:[`function_calling`]},{id:`meta-llama/llama-4-maverick`,name:`Meta: Llama 4 Maverick`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-4`,description:`Llama 4 Maverick 17B Instruct (128E) is a high-capacity multimodal language model from Meta, built on a mixture-of-experts (MoE) architecture with 128 experts and 17 billion active parameters per forward pass (400B total).`,context_window:1048576,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.15,output:.6},model_type:`chat`,release_date:`2025-04-05`,tagline:`Llama 4 Maverick 17B Instruct (128E) is a high-capacity multimodal language model from Meta, built on a mixture-of-experts (MoE) architecture with 128 experts and 17 billion active parameters per forw`,tools:[`function_calling`]},{id:`meta-llama/llama-4-scout`,name:`Meta: Llama 4 Scout`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-4`,description:`Llama 4 Scout 17B Instruct (16E) is a mixture-of-experts (MoE) language model developed by Meta, activating 17 billion parameters out of a total of 109B.`,context_window:327680,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.08,output:.3},model_type:`chat`,release_date:`2025-04-05`,tagline:`Llama 4 Scout 17B Instruct (16E) is a mixture-of-experts (MoE) language model developed by Meta, activating 17 billion parameters out of a total of 109B.`,tools:[`function_calling`]},{id:`meta-llama/llama-guard-3-8b`,name:`Llama Guard 3 8B`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-guard`,description:`Llama Guard 3 is a Llama-3.1-8B pretrained model, fine-tuned for content safety classification.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},pricing:{input:.02,output:.06},model_type:`moderation`,parameters:8,release_date:`2025-02-12`,tagline:`Llama Guard 3 is a Llama-3.1-8B pretrained model, fine-tuned for content safety classification.`},{id:`meta-llama/llama-guard-4-12b`,name:`Meta: Llama Guard 4 12B`,created_by:`meta-llama`,source:`official`,last_updated:`2026-03-23`,family:`llama-guard`,description:`Llama Guard 4 is a Llama 4 Scout-derived multimodal pretrained model, fine-tuned for content safety classification.`,context_window:163840,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.18,output:.18},model_type:`moderation`,parameters:12,release_date:`2025-04-30`,tagline:`Llama Guard 4 is a Llama 4 Scout-derived multimodal pretrained model, fine-tuned for content safety classification.`},{id:`microsoft/phi-4`,name:`Microsoft: Phi 4`,created_by:`microsoft`,source:`official`,last_updated:`2026-03-23`,description:`[Microsoft Research](/microsoft) Phi-4 is designed to perform well in complex reasoning tasks and can operate efficiently in situations with limited memory or where quick responses are needed.`,context_window:16384,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.065,output:.14},release_date:`2025-01-10`,tagline:`[Microsoft Research](/microsoft) Phi-4 is designed to perform well in complex reasoning tasks and can operate efficiently in situations with limited memory or where quick responses are needed.`},{id:`microsoft/wizardlm-2-8x22b`,name:`WizardLM-2 8x22B`,created_by:`microsoft`,source:`official`,last_updated:`2026-03-23`,description:`WizardLM-2 8x22B is Microsoft AI's most advanced Wizard model.`,context_window:65535,max_output_tokens:8e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:.62,output:.62},release_date:`2024-04-16`,tagline:`WizardLM-2 8x22B is Microsoft AI's most advanced Wizard model.`},{id:`minimax/minimax-01`,name:`MiniMax: MiniMax-01`,created_by:`minimax`,source:`official`,last_updated:`2026-03-23`,description:`MiniMax-01 is a combines MiniMax-Text-01 for text generation and MiniMax-VL-01 for image understanding.`,context_window:1000192,max_output_tokens:1000192,modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:.2,output:1.1},model_type:`chat`,family:`minimax`,release_date:`2025-01-15`,tagline:`MiniMax-01 is a combines MiniMax-Text-01 for text generation and MiniMax-VL-01 for image understanding.`},{id:`minimax/minimax-m1`,name:`MiniMax: MiniMax M1`,created_by:`minimax`,source:`official`,last_updated:`2026-03-23`,description:`MiniMax-M1 is a large-scale, open-weight reasoning model designed for extended context and high-efficiency inference.`,context_window:1e6,max_output_tokens:4e4,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:.4,output:2.2},model_type:`chat`,family:`minimax`,release_date:`2025-06-17`,tagline:`MiniMax-M1 is a large-scale, open-weight reasoning model designed for extended context and high-efficiency inference.`,tools:[`function_calling`]},{id:`minimax/minimax-m2-her`,name:`MiniMax: MiniMax M2-her`,created_by:`minimax`,source:`official`,last_updated:`2026-03-23`,description:`MiniMax M2-her is a dialogue-first large language model built for immersive roleplay, character-driven chat, and expressive multi-turn conversations.`,context_window:65536,max_output_tokens:2048,modalities:{input:[`text`],output:[`text`]},pricing:{input:.3,output:1.2,cached_input:.03},model_type:`chat`,family:`minimax`,release_date:`2026-01-23`,tagline:`MiniMax M2-her is a dialogue-first large language model built for immersive roleplay, character-driven chat, and expressive multi-turn conversations.`},{id:`minimax/minimax-m2.1`,name:`MiniMax: MiniMax M2.1`,created_by:`minimax`,source:`official`,last_updated:`2026-03-23`,description:`MiniMax-M2.1 is a lightweight, state-of-the-art large language model optimized for coding, agentic workflows, and modern application development.`,context_window:196608,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.27,output:.95,cached_input:.029},family:`minimax`,model_type:`chat`,release_date:`2025-12-23`,tagline:`MiniMax-M2.1 is a lightweight, state-of-the-art large language model optimized for coding, agentic workflows, and modern application development.`,tools:[`function_calling`],license:`mit`,open_weight:!0},{id:`minimax/minimax-m2.5:free`,name:`MiniMax: MiniMax M2.5 (free)`,created_by:`minimax`,source:`official`,last_updated:`2026-03-23`,description:`MiniMax-M2.5 is a SOTA large language model designed for real-world productivity.`,context_window:196608,max_output_tokens:196608,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},family:`minimax`,model_type:`chat`,release_date:`2026-02-12`,tagline:`MiniMax-M2.5 is a SOTA large language model designed for real-world productivity.`,tools:[`function_calling`]},{id:`minimax/minimax-m2.5`,name:`MiniMax: MiniMax M2.5`,created_by:`minimax`,source:`official`,last_updated:`2026-03-23`,description:`MiniMax-M2.5 is a SOTA large language model designed for real-world productivity.`,context_window:196608,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.2,output:1.17,cached_input:.1},family:`minimax`,model_type:`chat`,release_date:`2026-02-12`,tagline:`MiniMax-M2.5 is a SOTA large language model designed for real-world productivity.`,tools:[`function_calling`],license:`mit`,open_weight:!0},{id:`minimax/minimax-m2.7`,name:`MiniMax: MiniMax M2.7`,created_by:`minimax`,source:`official`,last_updated:`2026-03-23`,description:`MiniMax-M2.7 is a next-generation large language model designed for autonomous, real-world productivity and continuous improvement.`,context_window:204800,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.3,output:1.2,cached_input:.06},family:`minimax`,model_type:`chat`,release_date:`2026-03-18`,tagline:`MiniMax-M2.7 is a next-generation large language model designed for autonomous, real-world productivity and continuous improvement.`,tools:[`function_calling`],license:`proprietary`,open_weight:!1},{id:`minimax/minimax-m2`,name:`MiniMax: MiniMax M2`,created_by:`minimax`,source:`official`,last_updated:`2026-03-23`,description:`MiniMax-M2 is a compact, high-efficiency large language model optimized for end-to-end coding and agentic workflows.`,context_window:196608,max_output_tokens:196608,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.255,output:1,cached_input:.03},model_type:`chat`,family:`minimax`,release_date:`2025-10-23`,tagline:`MiniMax-M2 is a compact, high-efficiency large language model optimized for end-to-end coding and agentic workflows.`,tools:[`function_calling`],license:`proprietary`,open_weight:!1},{id:`mistralai/codestral-2508`,name:`Mistral: Codestral 2508`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`codestral`,description:`Mistral's cutting-edge language model for coding released end of July 2025.`,context_window:256e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:.3,output:.9,cached_input:.03},model_type:`code`,release_date:`2025-08-01`,alias:`mistralai/codestral`,tagline:`Mistral's cutting-edge language model for coding released end of July 2025.`,tools:[`function_calling`],license:`mnpl`,open_weight:!0},{id:`mistralai/codestral`,name:`Mistral: Codestral 2508`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`codestral`,description:`Mistral's cutting-edge language model for coding released end of July 2025.`,release_date:`2025-08-01`,context_window:256e3,model_type:`code`,tagline:`Mistral's cutting-edge language model for coding released end of July 2025.`,capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},modalities:{input:[`text`],output:[`text`]},pricing:{input:.3,output:.9,cached_input:.03},tools:[`function_calling`],snapshots:[`mistralai/codestral-2508`],license:`mnpl`,open_weight:!0},{id:`mistralai/devstral-2512`,name:`Mistral: Devstral 2 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`devstral`,description:`Devstral 2 is a state-of-the-art open-source model by Mistral AI specializing in agentic coding.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:2,cached_input:.04},model_type:`code`,release_date:`2025-12-09`,alias:`mistralai/devstral`,tagline:`Devstral 2 is a state-of-the-art open-source model by Mistral AI specializing in agentic coding.`,tools:[`function_calling`]},{id:`mistralai/devstral-medium`,name:`Mistral: Devstral Medium`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`devstral`,description:`Devstral Medium is a high-performance code generation and agentic reasoning model developed jointly by Mistral AI and All Hands AI.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:.4,output:2,cached_input:.04},model_type:`chat`,release_date:`2025-07-10`,tagline:`Devstral Medium is a high-performance code generation and agentic reasoning model developed jointly by Mistral AI and All Hands AI.`,tools:[`function_calling`],license:`proprietary`,open_weight:!1},{id:`mistralai/devstral-small`,name:`Mistral: Devstral Small 1.1`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`devstral`,description:`Devstral Small 1.1 is a 24B parameter open-weight language model for software engineering agents, developed by Mistral AI in collaboration with All Hands AI.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:.1,output:.3,cached_input:.01},model_type:`chat`,release_date:`2025-07-10`,tagline:`Devstral Small 1.1 is a 24B parameter open-weight language model for software engineering agents, developed by Mistral AI in collaboration with All Hands AI.`,tools:[`function_calling`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/devstral`,name:`Mistral: Devstral 2 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`devstral`,description:`Devstral 2 is a state-of-the-art open-source model by Mistral AI specializing in agentic coding.`,release_date:`2025-12-09`,context_window:262144,model_type:`code`,tagline:`Devstral 2 is a state-of-the-art open-source model by Mistral AI specializing in agentic coding.`,capabilities:{tool_call:!0,structured_output:!0},modalities:{input:[`text`],output:[`text`]},pricing:{input:.4,output:2,cached_input:.04},tools:[`function_calling`],snapshots:[`mistralai/devstral-2512`]},{id:`mistralai/ministral-14b-2512`,name:`Mistral: Ministral 3 14B 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`ministral`,description:`The largest model in the Ministral 3 family, Ministral 3 14B offers frontier capabilities and performance comparable to its larger Mistral Small 3.2 24B counterpart.`,context_window:262144,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:.2,output:.2,cached_input:.02},model_type:`chat`,parameters:14,release_date:`2025-12-02`,alias:`mistralai/ministral-14b`,tagline:`The largest model in the Ministral 3 family, Ministral 3 14B offers frontier capabilities and performance comparable to its larger Mistral Small 3.2 24B counterpart.`,tools:[`function_calling`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/ministral-14b`,name:`Mistral: Ministral 3 14B 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`ministral`,description:`The largest model in the Ministral 3 family, Ministral 3 14B offers frontier capabilities and performance comparable to its larger Mistral Small 3.2 24B counterpart.`,release_date:`2025-12-02`,context_window:262144,parameters:14,model_type:`chat`,tagline:`The largest model in the Ministral 3 family, Ministral 3 14B offers frontier capabilities and performance comparable to its larger Mistral Small 3.2 24B counterpart.`,capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:.2,output:.2,cached_input:.02},tools:[`function_calling`],snapshots:[`mistralai/ministral-14b-2512`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/ministral-3b-2512`,name:`Mistral: Ministral 3 3B 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`ministral`,description:`The smallest model in the Ministral 3 family, Ministral 3 3B is a powerful, efficient tiny language model with vision capabilities.`,context_window:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:.1,output:.1,cached_input:.01},model_type:`chat`,parameters:3,release_date:`2025-12-02`,alias:`mistralai/ministral-3b`,tagline:`The smallest model in the Ministral 3 family, Ministral 3 3B is a powerful, efficient tiny language model with vision capabilities.`,tools:[`function_calling`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/ministral-3b`,name:`Mistral: Ministral 3 3B 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`ministral`,description:`The smallest model in the Ministral 3 family, Ministral 3 3B is a powerful, efficient tiny language model with vision capabilities.`,release_date:`2025-12-02`,context_window:131072,parameters:3,model_type:`chat`,tagline:`The smallest model in the Ministral 3 family, Ministral 3 3B is a powerful, efficient tiny language model with vision capabilities.`,capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:.1,output:.1,cached_input:.01},tools:[`function_calling`],snapshots:[`mistralai/ministral-3b-2512`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/ministral-8b-2512`,name:`Mistral: Ministral 3 8B 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`ministral`,description:`A balanced model in the Ministral 3 family, Ministral 3 8B is a powerful, efficient tiny language model with vision capabilities.`,context_window:262144,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:.15,output:.15,cached_input:.015},model_type:`chat`,parameters:8,release_date:`2025-12-02`,alias:`mistralai/ministral-8b`,tagline:`A balanced model in the Ministral 3 family, Ministral 3 8B is a powerful, efficient tiny language model with vision capabilities.`,tools:[`function_calling`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/ministral-8b`,name:`Mistral: Ministral 3 8B 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`ministral`,description:`A balanced model in the Ministral 3 family, Ministral 3 8B is a powerful, efficient tiny language model with vision capabilities.`,release_date:`2025-12-02`,context_window:262144,parameters:8,model_type:`chat`,tagline:`A balanced model in the Ministral 3 family, Ministral 3 8B is a powerful, efficient tiny language model with vision capabilities.`,capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:.15,output:.15,cached_input:.015},tools:[`function_calling`],snapshots:[`mistralai/ministral-8b-2512`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/mistral-7b-instruct-v0.1`,name:`Mistral: Mistral 7B Instruct v0.1`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,description:`A 7.3B parameter model that outperforms Llama 2 13B on all benchmarks, with optimizations for speed and context length.`,context_window:2824,modalities:{input:[`text`],output:[`text`]},pricing:{input:.11,output:.19},model_type:`chat`,family:`mistral-7b`,parameters:7,release_date:`2023-09-28`,tagline:`A 7.3B parameter model that outperforms Llama 2 13B on all benchmarks, with optimizations for speed and context length.`},{id:`mistralai/mistral-large-2407`,name:`Mistral Large 2407`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-large`,description:`This is Mistral AI's flagship model, Mistral Large 2 (version mistral-large-2407).`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:2,output:6,cached_input:.2},model_type:`chat`,release_date:`2024-11-19`,alias:`mistralai/mistral-large`,tagline:`This is Mistral AI's flagship model, Mistral Large 2 (version mistral-large-2407).`,tools:[`function_calling`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/mistral-large-2411`,name:`Mistral Large 2411`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-large`,description:`Mistral Large 2 2411 is an update of [Mistral Large 2](/mistralai/mistral-large) released together with [Pixtral Large 2411](/mistralai/pixtral-large-2411)
+
+It provides a significant upgrade on the pr`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:2,output:6,cached_input:.2},model_type:`chat`,release_date:`2024-11-19`,alias:`mistralai/mistral-large`,tagline:`Mistral Large 2 2411 is an update of [Mistral Large 2](/mistralai/mistral-large) released together with [Pixtral Large 2411](/mistralai/pixtral-large-2411)
+
+It provides a significant upgrade on the pr`,tools:[`function_calling`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/mistral-large-2512`,name:`Mistral: Mistral Large 3 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-large`,description:`Mistral Large 3 2512 is Mistral’s most capable model to date, featuring a sparse mixture-of-experts architecture with 41B active parameters (675B total), and released under the Apache 2.0 license.`,context_window:262144,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:.5,output:1.5,cached_input:.05},model_type:`chat`,release_date:`2025-12-01`,alias:`mistralai/mistral-large`,tagline:`Mistral Large 3 2512 is Mistral’s most capable model to date, featuring a sparse mixture-of-experts architecture with 41B active parameters (675B total), and released under the Apache 2.0 license.`,tools:[`function_calling`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/mistral-large`,name:`Mistral Large`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-large`,description:"This is Mistral AI's flagship model, Mistral Large 2 (version `mistral-large-2407`).",context_window:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:2,output:6,cached_input:.2},model_type:`chat`,release_date:`2024-02-26`,tagline:`Mistral Large 3 2512 is Mistral’s most capable model to date, featuring a sparse mixture-of-experts architecture with 41B active parameters (675B total), and released under the Apache 2.0 license.`,tools:[`function_calling`],snapshots:[`mistralai/mistral-large-2512`,`mistralai/mistral-large-2411`,`mistralai/mistral-large-2407`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/mistral-medium-3.1`,name:`Mistral: Mistral Medium 3.1`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-medium`,description:`Mistral Medium 3.1 is an updated version of Mistral Medium 3, which is a high-performance enterprise-grade language model designed to deliver frontier-level capabilities at significantly reduced operational cost.`,context_window:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:2,cached_input:.04},model_type:`chat`,release_date:`2025-08-13`,tagline:`Mistral Medium 3.1 is an updated version of Mistral Medium 3, which is a high-performance enterprise-grade language model designed to deliver frontier-level capabilities at significantly reduced opera`,tools:[`function_calling`]},{id:`mistralai/mistral-medium-3`,name:`Mistral: Mistral Medium 3`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-medium`,description:`Mistral Medium 3 is a high-performance enterprise-grade language model designed to deliver frontier-level capabilities at significantly reduced operational cost.`,context_window:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:2,cached_input:.04},model_type:`chat`,release_date:`2025-05-07`,tagline:`Mistral Medium 3 is a high-performance enterprise-grade language model designed to deliver frontier-level capabilities at significantly reduced operational cost.`,tools:[`function_calling`]},{id:`mistralai/mistral-nemo`,name:`Mistral: Mistral Nemo`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,description:`A 12B parameter model with a 128k token context length built by Mistral in collaboration with NVIDIA.`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:.02,output:.04},model_type:`chat`,family:`mistral-nemo`,release_date:`2024-07-19`,tagline:`A 12B parameter model with a 128k token context length built by Mistral in collaboration with NVIDIA.`,tools:[`function_calling`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/mistral-saba`,name:`Mistral: Saba`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,description:`Mistral Saba is a 24B-parameter language model specifically designed for the Middle East and South Asia, delivering accurate and contextually relevant responses while maintaining efficient performance.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:.2,output:.6,cached_input:.02},model_type:`chat`,family:`mistral-saba`,release_date:`2025-02-17`,tagline:`Mistral Saba is a 24B-parameter language model specifically designed for the Middle East and South Asia, delivering accurate and contextually relevant responses while maintaining efficient performance`,tools:[`function_calling`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/mistral-small-24b-instruct-2501`,name:`Mistral: Mistral Small 3`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-small`,description:`Mistral Small 3 is a 24B-parameter language model optimized for low-latency performance across common AI tasks.`,context_window:32768,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.05,output:.08},model_type:`chat`,parameters:24,release_date:`2025-01-30`,alias:`mistralai/mistral-small-24b-instruct`,tagline:`Mistral Small 3 is a 24B-parameter language model optimized for low-latency performance across common AI tasks.`,tools:[`function_calling`]},{id:`mistralai/mistral-small-24b-instruct`,name:`Mistral: Mistral Small 3`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-small`,description:`Mistral Small 3 is a 24B-parameter language model optimized for low-latency performance across common AI tasks.`,release_date:`2025-01-30`,context_window:32768,max_output_tokens:16384,parameters:24,model_type:`chat`,tagline:`Mistral Small 3 is a 24B-parameter language model optimized for low-latency performance across common AI tasks.`,capabilities:{tool_call:!0,structured_output:!0},modalities:{input:[`text`],output:[`text`]},pricing:{input:.05,output:.08},tools:[`function_calling`],snapshots:[`mistralai/mistral-small-24b-instruct-2501`]},{id:`mistralai/mistral-small-2603`,name:`Mistral: Mistral Small 4`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-small`,description:`Mistral Small 4 is the next major release in the Mistral Small family, unifying the capabilities of several flagship Mistral models into a single system.`,context_window:262144,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:.15,output:.6,cached_input:.015},model_type:`chat`,release_date:`2026-03-16`,alias:`mistralai/mistral-small`,tagline:`Mistral Small 4 is the next major release in the Mistral Small family, unifying the capabilities of several flagship Mistral models into a single system.`,tools:[`function_calling`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/mistral-small-3.1-24b-instruct:free`,name:`Mistral: Mistral Small 3.1 24B (free)`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-small`,description:`Mistral Small 3.1 24B Instruct is an upgraded variant of Mistral Small 3 (2501), featuring 24 billion parameters with advanced multimodal capabilities.`,status:`deprecated`,deprecation_date:`2026-03-29`,context_window:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},model_type:`chat`,parameters:24,release_date:`2025-03-17`,tagline:`Mistral Small 3.1 24B Instruct is an upgraded variant of Mistral Small 3 (2501), featuring 24 billion parameters with advanced multimodal capabilities.`,tools:[`function_calling`]},{id:`mistralai/mistral-small-3.1-24b-instruct`,name:`Mistral: Mistral Small 3.1 24B`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-small`,description:`Mistral Small 3.1 24B Instruct is an upgraded variant of Mistral Small 3 (2501), featuring 24 billion parameters with advanced multimodal capabilities.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.03,output:.11,cached_input:.015},model_type:`chat`,parameters:24,release_date:`2025-03-17`,tagline:`Mistral Small 3.1 24B Instruct is an upgraded variant of Mistral Small 3 (2501), featuring 24 billion parameters with advanced multimodal capabilities.`},{id:`mistralai/mistral-small-3.2-24b-instruct`,name:`Mistral: Mistral Small 3.2 24B`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-small`,description:`Mistral-Small-3.2-24B-Instruct-2506 is an updated 24B parameter model from Mistral optimized for instruction following, repetition reduction, and improved function calling.`,context_window:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.075,output:.2},model_type:`chat`,parameters:24,release_date:`2025-06-20`,tagline:`Mistral-Small-3.2-24B-Instruct-2506 is an updated 24B parameter model from Mistral optimized for instruction following, repetition reduction, and improved function calling.`,tools:[`function_calling`]},{id:`mistralai/mistral-small-creative`,name:`Mistral: Mistral Small Creative`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-small`,description:`Mistral Small Creative is an experimental small model designed for creative writing, narrative generation, roleplay and character-driven dialogue, general-purpose instruction following, and conversati`,status:`deprecated`,deprecation_date:`2026-04-30`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,streaming:!0,vision:!0},pricing:{input:.1,output:.3,cached_input:.01},model_type:`chat`,release_date:`2025-12-16`,tagline:`Mistral Small Creative is an experimental small model designed for creative writing, narrative generation, roleplay and character-driven dialogue, general-purpose instruction following, and conversati`,tools:[`function_calling`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/mistral-small`,name:`Mistral: Mistral Small 4`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mistral-small`,description:`Mistral Small 4 is the next major release in the Mistral Small family, unifying the capabilities of several flagship Mistral models into a single system.`,release_date:`2026-03-16`,context_window:262144,model_type:`chat`,tagline:`Mistral Small 4 is the next major release in the Mistral Small family, unifying the capabilities of several flagship Mistral models into a single system.`,capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:.15,output:.6,cached_input:.015},tools:[`function_calling`],snapshots:[`mistralai/mistral-small-2603`],license:`apache-2.0`,open_weight:!0},{id:`mistralai/mixtral-8x22b-instruct`,name:`Mistral: Mixtral 8x22B Instruct`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mixtral`,description:`Mistral's official instruct fine-tuned version of [Mixtral 8x22B](/models/mistralai/mixtral-8x22b).`,context_window:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2,output:6,cached_input:.2},model_type:`chat`,release_date:`2024-04-17`,tagline:`Mistral's official instruct fine-tuned version of [Mixtral 8x22B](/models/mistralai/mixtral-8x22b).`,tools:[`function_calling`]},{id:`mistralai/mixtral-8x7b-instruct`,name:`Mistral: Mixtral 8x7B Instruct`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`mixtral`,description:`Mixtral 8x7B Instruct is a pretrained generative Sparse Mixture of Experts, by Mistral AI, for chat and instruction use.`,context_window:32768,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.54,output:.54},model_type:`chat`,release_date:`2023-12-10`,tagline:`Mixtral 8x7B Instruct is a pretrained generative Sparse Mixture of Experts, by Mistral AI, for chat and instruction use.`,tools:[`function_calling`]},{id:`mistralai/pixtral-12b`,name:`Mistral: Pixtral 12B`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`pixtral`,description:`The first multi-modal, text+image-to-text model from Mistral AI.`,context_window:32768,modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:.1,output:.1},model_type:`chat`,parameters:12,release_date:`2024-09-10`,tagline:`The first multi-modal, text+image-to-text model from Mistral AI.`,license:`apache-2.0`,open_weight:!0,capabilities:{streaming:!0,vision:!0}},{id:`mistralai/pixtral-large-2411`,name:`Mistral: Pixtral Large 2411`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`pixtral`,description:`Pixtral Large is a 124B parameter, open-weight, multimodal model built on top of [Mistral Large 2](/mistralai/mistral-large-2411).`,context_window:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:2,output:6,cached_input:.2},model_type:`chat`,release_date:`2024-11-19`,alias:`mistralai/pixtral-large`,tagline:`Pixtral Large is a 124B parameter, open-weight, multimodal model built on top of [Mistral Large 2](/mistralai/mistral-large-2411).`,tools:[`function_calling`],license:`mnpl`,open_weight:!0},{id:`mistralai/pixtral-large`,name:`Mistral: Pixtral Large 2411`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`pixtral`,description:`Pixtral Large is a 124B parameter, open-weight, multimodal model built on top of [Mistral Large 2](/mistralai/mistral-large-2411).`,release_date:`2024-11-19`,context_window:131072,model_type:`chat`,tagline:`Pixtral Large is a 124B parameter, open-weight, multimodal model built on top of [Mistral Large 2](/mistralai/mistral-large-2411).`,capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:2,output:6,cached_input:.2},tools:[`function_calling`],snapshots:[`mistralai/pixtral-large-2411`],license:`mnpl`,open_weight:!0},{id:`mistralai/voxtral-small-24b-2507`,name:`Mistral: Voxtral Small 24B 2507`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,description:`Voxtral Small is an enhancement of Mistral Small 3, incorporating state-of-the-art audio input capabilities while retaining best-in-class text performance.`,context_window:32e3,modalities:{input:[`text`,`audio`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.1,output:.3,cached_input:.01},model_type:`audio`,family:`voxtral`,parameters:24,release_date:`2025-10-30`,alias:`mistralai/voxtral-small-24b`,tagline:`Voxtral Small is an enhancement of Mistral Small 3, incorporating state-of-the-art audio input capabilities while retaining best-in-class text performance.`,tools:[`function_calling`]},{id:`mistralai/voxtral-small-24b`,name:`Mistral: Voxtral Small 24B 2507`,created_by:`mistral`,source:`official`,last_updated:`2026-03-23`,family:`voxtral`,description:`Voxtral Small is an enhancement of Mistral Small 3, incorporating state-of-the-art audio input capabilities while retaining best-in-class text performance.`,release_date:`2025-10-30`,context_window:32e3,parameters:24,model_type:`audio`,tagline:`Voxtral Small is an enhancement of Mistral Small 3, incorporating state-of-the-art audio input capabilities while retaining best-in-class text performance.`,capabilities:{tool_call:!0,structured_output:!0},modalities:{input:[`text`,`audio`],output:[`text`]},pricing:{input:.1,output:.3,cached_input:.01},tools:[`function_calling`],snapshots:[`mistralai/voxtral-small-24b-2507`]},{id:`moonshotai/kimi-k2-0905`,name:`MoonshotAI: Kimi K2 0905`,created_by:`moonshotai`,source:`official`,last_updated:`2026-03-23`,description:`Kimi K2 0905 is the September update of [Kimi K2 0711](moonshotai/kimi-k2).`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:2,cached_input:.15},family:`kimi-k2`,model_type:`chat`,release_date:`2025-09-04`,tagline:`Kimi K2 0905 is the September update of [Kimi K2 0711](moonshotai/kimi-k2).`,tools:[`function_calling`]},{id:`moonshotai/kimi-k2-thinking`,name:`MoonshotAI: Kimi K2 Thinking`,created_by:`moonshotai`,source:`official`,last_updated:`2026-03-23`,description:`Kimi K2 Thinking is Moonshot AI’s most advanced open reasoning model to date, extending the K2 series into agentic, long-horizon reasoning.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.47,output:2,cached_input:.141},family:`kimi-k2`,model_type:`chat`,release_date:`2025-11-06`,tagline:`Kimi K2 Thinking is Moonshot AI’s most advanced open reasoning model to date, extending the K2 series into agentic, long-horizon reasoning.`,tools:[`function_calling`]},{id:`moonshotai/kimi-k2.5`,name:`MoonshotAI: Kimi K2.5`,created_by:`moonshotai`,source:`official`,last_updated:`2026-03-23`,description:`Kimi K2.5 is Moonshot AI's native multimodal model, delivering state-of-the-art visual coding capability and a self-directed agent swarm paradigm.`,context_window:262144,max_output_tokens:65535,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.45,output:2.2,cached_input:.225},family:`kimi-k2.5`,model_type:`chat`,release_date:`2026-01-27`,tagline:`Kimi K2.5 is Moonshot AI's native multimodal model, delivering state-of-the-art visual coding capability and a self-directed agent swarm paradigm.`,tools:[`function_calling`]},{id:`moonshotai/kimi-k2`,name:`MoonshotAI: Kimi K2 0711`,created_by:`moonshotai`,source:`official`,last_updated:`2026-03-23`,description:`Kimi K2 Instruct is a large-scale Mixture-of-Experts (MoE) language model developed by Moonshot AI, featuring 1 trillion total parameters with 32 billion active per forward pass.`,context_window:131e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.55,output:2.2},family:`kimi-k2`,model_type:`chat`,release_date:`2025-07-11`,tagline:`Kimi K2 Instruct is a large-scale Mixture-of-Experts (MoE) language model developed by Moonshot AI, featuring 1 trillion total parameters with 32 billion active per forward pass.`,tools:[`function_calling`]},{id:`morph/morph-v3-fast`,name:`Morph: Morph V3 Fast`,created_by:`morph`,source:`official`,last_updated:`2026-03-23`,description:`Morph's fastest apply model for code edits.`,context_window:81920,max_output_tokens:38e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:.8,output:1.2},release_date:`2025-07-07`,tagline:`Morph's fastest apply model for code edits.`},{id:`morph/morph-v3-large`,name:`Morph: Morph V3 Large`,created_by:`morph`,source:`official`,last_updated:`2026-03-23`,description:`Morph's high-accuracy apply model for complex code edits.`,context_window:262144,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},pricing:{input:.9,output:1.9},release_date:`2025-07-07`,tagline:`Morph's high-accuracy apply model for complex code edits.`},{id:`nex-agi/deepseek-v3.1-nex-n1`,name:`Nex AGI: DeepSeek V3.1 Nex N1`,created_by:`nex-agi`,source:`official`,last_updated:`2026-03-23`,family:`deepseek`,description:`DeepSeek V3.1 Nex-N1 is the flagship release of the Nex-N1 series — a post-trained model designed to highlight agent autonomy, tool use, and real-world productivity.`,context_window:131072,max_output_tokens:163840,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.27,output:1},release_date:`2025-12-08`,tagline:`DeepSeek V3.1 Nex-N1 is the flagship release of the Nex-N1 series — a post-trained model designed to highlight agent autonomy, tool use, and real-world productivity.`,tools:[`function_calling`]},{id:`nousresearch/hermes-2-pro-llama-3-8b`,name:`NousResearch: Hermes 2 Pro - Llama-3 8B`,created_by:`nousresearch`,source:`official`,last_updated:`2026-03-23`,description:`Hermes 2 Pro is an upgraded, retrained version of Nous Hermes 2, consisting of an updated and cleaned version of the OpenHermes 2.5 Dataset, as well as a newly introduced Function Calling and JSON Mod`,context_window:8192,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.14,output:.14},model_type:`chat`,parameters:8,release_date:`2024-05-27`,tagline:`Hermes 2 Pro is an upgraded, retrained version of Nous Hermes 2, consisting of an updated and cleaned version of the OpenHermes 2.5 Dataset, as well as a newly introduced Function Calling and JSON Mod`},{id:`nousresearch/hermes-3-llama-3.1-405b:free`,name:`Nous: Hermes 3 405B Instruct (free)`,created_by:`nousresearch`,source:`official`,last_updated:`2026-03-23`,description:`Hermes 3 is a generalist language model with many improvements over Hermes 2, including advanced agentic capabilities, much better roleplaying, reasoning, multi-turn conversation, long context coherence, and improvements across the board.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},model_type:`chat`,release_date:`2024-08-16`,tagline:`Hermes 3 is a generalist language model with many improvements over Hermes 2, including advanced agentic capabilities, much better roleplaying, reasoning, multi-turn conversation, long context coheren`},{id:`nousresearch/hermes-3-llama-3.1-405b`,name:`Nous: Hermes 3 405B Instruct`,created_by:`nousresearch`,source:`official`,last_updated:`2026-03-23`,description:`Hermes 3 is a generalist language model with many improvements over Hermes 2, including advanced agentic capabilities, much better roleplaying, reasoning, multi-turn conversation, long context coherence, and improvements across the board.`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:1,output:1},model_type:`chat`,parameters:405,release_date:`2024-08-16`,tagline:`Hermes 3 is a generalist language model with many improvements over Hermes 2, including advanced agentic capabilities, much better roleplaying, reasoning, multi-turn conversation, long context coheren`},{id:`nousresearch/hermes-3-llama-3.1-70b`,name:`Nous: Hermes 3 70B Instruct`,created_by:`nousresearch`,source:`official`,last_updated:`2026-03-23`,description:`Hermes 3 is a generalist language model with many improvements over [Hermes 2](/models/nousresearch/nous-hermes-2-mistral-7b-dpo), including advanced agentic capabilities, much better roleplaying, reasoning, multi-turn conversation, long context coherence, and improvements across the board.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.3,output:.3},model_type:`chat`,parameters:70,release_date:`2024-08-18`,tagline:`Hermes 3 is a generalist language model with many improvements over [Hermes 2](/models/nousresearch/nous-hermes-2-mistral-7b-dpo), including advanced agentic capabilities, much better roleplaying, rea`},{id:`nousresearch/hermes-4-405b`,name:`Nous: Hermes 4 405B`,created_by:`nousresearch`,source:`official`,last_updated:`2026-03-23`,description:`Hermes 4 is a large-scale reasoning model built on Meta-Llama-3.1-405B and released by Nous Research.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:1,output:3},model_type:`chat`,parameters:405,release_date:`2025-08-26`,tagline:`Hermes 4 is a large-scale reasoning model built on Meta-Llama-3.1-405B and released by Nous Research.`},{id:`nousresearch/hermes-4-70b`,name:`Nous: Hermes 4 70B`,created_by:`nousresearch`,source:`official`,last_updated:`2026-03-23`,description:`Hermes 4 70B is a hybrid reasoning model from Nous Research, built on Meta-Llama-3.1-70B.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.13,output:.4},model_type:`chat`,parameters:70,release_date:`2025-08-26`,tagline:`Hermes 4 70B is a hybrid reasoning model from Nous Research, built on Meta-Llama-3.1-70B.`},{id:`nvidia/llama-3.1-nemotron-70b-instruct`,name:`NVIDIA: Llama 3.1 Nemotron 70B Instruct`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-23`,family:`llama-3.1`,description:`NVIDIA's Llama 3.1 Nemotron 70B is a language model designed for generating precise and useful responses.`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:1.2,output:1.2},model_type:`chat`,parameters:70,release_date:`2024-10-15`,tagline:`NVIDIA's Llama 3.1 Nemotron 70B is a language model designed for generating precise and useful responses.`,tools:[`function_calling`]},{id:`nvidia/llama-3.1-nemotron-ultra-253b-v1`,name:`NVIDIA: Llama 3.1 Nemotron Ultra 253B v1`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-23`,family:`llama-3.1`,description:`Llama-3.1-Nemotron-Ultra-253B-v1 is a large language model (LLM) optimized for advanced reasoning, human-interactive chat, retrieval-augmented generation (RAG), and tool-calling tasks.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.6,output:1.8},model_type:`chat`,parameters:253,release_date:`2025-04-08`,tagline:`Llama-3.1-Nemotron-Ultra-253B-v1 is a large language model (LLM) optimized for advanced reasoning, human-interactive chat, retrieval-augmented generation (RAG), and tool-calling tasks.`},{id:`nvidia/llama-3.3-nemotron-super-49b-v1.5`,name:`NVIDIA: Llama 3.3 Nemotron Super 49B V1.5`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-23`,family:`llama-3.3`,description:`Llama-3.3-Nemotron-Super-49B-v1.5 is a 49B-parameter, English-centric reasoning/chat model derived from Meta’s Llama-3.3-70B-Instruct with a 128K context.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.1,output:.4},model_type:`chat`,parameters:49,release_date:`2025-10-10`,tagline:`Llama-3.3-Nemotron-Super-49B-v1.5 is a 49B-parameter, English-centric reasoning/chat model derived from Meta’s Llama-3.3-70B-Instruct with a 128K context.`,tools:[`function_calling`]},{id:`nvidia/nemotron-3-nano-30b-a3b:free`,name:`NVIDIA: Nemotron 3 Nano 30B A3B (free)`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-23`,description:`NVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems.`,context_window:256e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},model_type:`chat`,parameters:30,active_parameters:3,release_date:`2025-12-14`,tagline:`NVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems.`,tools:[`function_calling`]},{id:`nvidia/nemotron-3-nano-30b-a3b`,name:`NVIDIA: Nemotron 3 Nano 30B A3B`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-23`,description:`NVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.05,output:.2},model_type:`chat`,parameters:30,active_parameters:3,release_date:`2025-12-14`,tagline:`NVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems.`,tools:[`function_calling`]},{id:`nvidia/nemotron-3-super-120b-a12b:free`,name:`NVIDIA: Nemotron 3 Super (free)`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-23`,description:`NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications.`,context_window:262144,max_output_tokens:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},model_type:`chat`,parameters:120,active_parameters:12,release_date:`2026-03-11`,tagline:`NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications.`,tools:[`function_calling`]},{id:`nvidia/nemotron-3-super-120b-a12b`,name:`NVIDIA: Nemotron 3 Super`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-23`,description:`NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.1,output:.5,cached_input:.04},model_type:`chat`,parameters:120,active_parameters:12,release_date:`2026-03-11`,tagline:`NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications.`,tools:[`function_calling`]},{id:`nvidia/nemotron-nano-12b-v2-vl:free`,name:`NVIDIA: Nemotron Nano 12B 2 VL (free)`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-23`,description:`NVIDIA Nemotron Nano 2 VL is a 12-billion-parameter open multimodal reasoning model designed for video understanding and document intelligence.`,context_window:128e3,max_output_tokens:128e3,modalities:{input:[`image`,`text`,`video`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},model_type:`chat`,parameters:12,release_date:`2025-10-28`,tagline:`NVIDIA Nemotron Nano 2 VL is a 12-billion-parameter open multimodal reasoning model designed for video understanding and document intelligence.`,tools:[`function_calling`]},{id:`nvidia/nemotron-nano-12b-v2-vl`,name:`NVIDIA: Nemotron Nano 12B 2 VL`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-23`,description:`NVIDIA Nemotron Nano 2 VL is a 12-billion-parameter open multimodal reasoning model designed for video understanding and document intelligence.`,context_window:131072,modalities:{input:[`image`,`text`,`video`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.2,output:.6},model_type:`chat`,parameters:12,release_date:`2025-10-28`,tagline:`NVIDIA Nemotron Nano 2 VL is a 12-billion-parameter open multimodal reasoning model designed for video understanding and document intelligence.`},{id:`nvidia/nemotron-nano-9b-v2:free`,name:`NVIDIA: Nemotron Nano 9B V2 (free)`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-23`,description:`NVIDIA-Nemotron-Nano-9B-v2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},model_type:`chat`,parameters:9,release_date:`2025-09-05`,tagline:`NVIDIA-Nemotron-Nano-9B-v2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks.`,tools:[`function_calling`]},{id:`nvidia/nemotron-nano-9b-v2`,name:`NVIDIA: Nemotron Nano 9B V2`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-23`,description:`NVIDIA-Nemotron-Nano-9B-v2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.04,output:.16},model_type:`chat`,parameters:9,release_date:`2025-09-05`,tagline:`NVIDIA-Nemotron-Nano-9B-v2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks.`,tools:[`function_calling`]},{id:`openai/gpt-3.5-turbo-0613`,name:`OpenAI: GPT-3.5 Turbo (older v0613)`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-3.5`,description:`GPT-3.5 Turbo is OpenAI's fastest model.`,context_window:4095,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,fine_tuning:!0},pricing:{input:1,output:2},model_type:`chat`,release_date:`2024-01-25`,tagline:`GPT-3.5 Turbo is OpenAI's fastest model.`,tools:[`function_calling`],knowledge_cutoff:`2021-08`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-3.5-turbo-16k`,name:`OpenAI: GPT-3.5 Turbo 16k`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-3.5`,description:`This model offers four times the context length of gpt-3.5-turbo, allowing it to support approximately 20 pages of text in a single request at a higher cost.`,context_window:16385,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:3,output:4},model_type:`chat`,release_date:`2023-08-28`,tagline:`This model offers four times the context length of gpt-3.5-turbo, allowing it to support approximately 20 pages of text in a single request at a higher cost.`,tools:[`function_calling`]},{id:`openai/gpt-3.5-turbo-instruct`,name:`OpenAI: GPT-3.5 Turbo Instruct`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-3.5`,description:`This model is a variant of GPT-3.5 Turbo tuned for instructional prompts and omitting chat-related optimizations.`,context_window:4095,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,fine_tuning:!0},pricing:{input:1.5,output:2},model_type:`chat`,release_date:`2023-09-28`,tagline:`This model is a variant of GPT-3.5 Turbo tuned for instructional prompts and omitting chat-related optimizations.`,knowledge_cutoff:`2021-08`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-3.5-turbo`,name:`OpenAI: GPT-3.5 Turbo`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-3.5`,description:`GPT-3.5 Turbo is OpenAI's fastest model.`,context_window:16385,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,fine_tuning:!0},pricing:{input:.5,output:1.5},model_type:`chat`,release_date:`2023-05-28`,tagline:`GPT-3.5 Turbo is OpenAI's fastest model.`,tools:[`function_calling`],knowledge_cutoff:`2021-08`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4-0314`,name:`OpenAI: GPT-4 (older v0314)`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4`,description:`GPT-4-0314 is the first version of GPT-4 released, with a context length of 8,192 tokens, and was supported until June 14.`,context_window:8191,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:30,output:60},model_type:`chat`,release_date:`2023-05-28`,tagline:`GPT-4-0314 is the first version of GPT-4 released, with a context length of 8,192 tokens, and was supported until June 14.`,tools:[`function_calling`]},{id:`openai/gpt-4-1106-preview`,name:`OpenAI: GPT-4 Turbo (older v1106)`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4`,description:`The latest GPT-4 Turbo model with vision capabilities.`,context_window:128e3,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:10,output:30},model_type:`chat`,release_date:`2023-11-06`,tagline:`The latest GPT-4 Turbo model with vision capabilities.`,tools:[`function_calling`]},{id:`openai/gpt-4-turbo-preview`,name:`OpenAI: GPT-4 Turbo Preview`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4-turbo`,description:`The preview GPT-4 model with improved instruction following, JSON mode, reproducible outputs, parallel function calling, and more.`,context_window:128e3,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:10,output:30},model_type:`chat`,release_date:`2024-01-25`,tagline:`The preview GPT-4 model with improved instruction following, JSON mode, reproducible outputs, parallel function calling, and more.`,tools:[`function_calling`],license:`proprietary`,open_weight:!1},{id:`openai/gpt-4-turbo`,name:`OpenAI: GPT-4 Turbo`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4-turbo`,description:`The latest GPT-4 Turbo model with vision capabilities.`,context_window:128e3,max_output_tokens:4096,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:10,output:30},model_type:`chat`,release_date:`2024-04-09`,tagline:`The latest GPT-4 Turbo model with vision capabilities.`,tools:[`function_calling`],knowledge_cutoff:`2023-11`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4.1-mini`,name:`OpenAI: GPT-4.1 Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4.1`,description:`GPT-4.1 Mini is a mid-sized model delivering performance competitive with GPT-4o at substantially lower latency and cost.`,context_window:1047576,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0,fine_tuning:!0},pricing:{input:.4,output:1.6,cached_input:.1},model_type:`chat`,release_date:`2025-04-14`,tagline:`GPT-4.1 Mini is a mid-sized model delivering performance competitive with GPT-4o at substantially lower latency and cost.`,tools:[`function_calling`],knowledge_cutoff:`2024-05`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4.1-nano`,name:`OpenAI: GPT-4.1 Nano`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4.1`,description:`For tasks that demand low latency, GPT‑4.1 nano is the fastest and cheapest model in the GPT-4.1 series.`,context_window:1047576,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0,fine_tuning:!0},pricing:{input:.1,output:.4,cached_input:.025},model_type:`chat`,release_date:`2025-04-14`,tagline:`For tasks that demand low latency, GPT‑4.1 nano is the fastest and cheapest model in the GPT-4.1 series.`,tools:[`function_calling`],knowledge_cutoff:`2024-05`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4.1`,name:`OpenAI: GPT-4.1`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4.1`,description:`GPT-4.1 is a flagship large language model optimized for advanced instruction following, real-world software engineering, and long-context reasoning.`,context_window:1047576,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0,fine_tuning:!0},pricing:{input:2,output:8,cached_input:.5},model_type:`chat`,release_date:`2025-04-14`,tagline:`GPT-4.1 is a flagship large language model optimized for advanced instruction following, real-world software engineering, and long-context reasoning.`,tools:[`function_calling`],knowledge_cutoff:`2024-05`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4`,name:`OpenAI: GPT-4`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4`,description:`OpenAI's flagship model, GPT-4 is a large-scale multimodal language model capable of solving difficult problems with greater accuracy than previous models due to its broader general knowledge and advanced reasoning capabilities.`,context_window:8191,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,fine_tuning:!0},pricing:{input:30,output:60},model_type:`chat`,release_date:`2023-05-28`,tagline:`OpenAI's flagship model, GPT-4 is a large-scale multimodal language model capable of solving difficult problems with greater accuracy than previous models due to its broader general knowledge and adva`,tools:[`function_calling`],knowledge_cutoff:`2023-11`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4o-2024-05-13`,name:`OpenAI: GPT-4o (2024-05-13)`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4o`,description:`GPT-4o ("o" for "omni") is OpenAI's latest AI model, supporting both text and image inputs with text outputs.`,context_window:128e3,max_output_tokens:4096,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0,fine_tuning:!0},pricing:{input:5,output:15,cached_input:1.25},model_type:`chat`,release_date:`2024-05-13`,alias:`openai/gpt-4o`,tagline:`GPT-4o ("o" for "omni") is OpenAI's latest AI model, supporting both text and image inputs with text outputs.`,tools:[`function_calling`],knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4o-2024-08-06`,name:`OpenAI: GPT-4o (2024-08-06)`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4o`,description:`The 2024-08-06 version of GPT-4o offers improved performance in structured outputs, with the ability to supply a JSON schema in the respone_format.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0,fine_tuning:!0},pricing:{input:2.5,output:10,cached_input:1.25},model_type:`chat`,release_date:`2024-08-06`,alias:`openai/gpt-4o`,tagline:`The 2024-08-06 version of GPT-4o offers improved performance in structured outputs, with the ability to supply a JSON schema in the respone_format.`,tools:[`function_calling`],knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4o-2024-11-20`,name:`OpenAI: GPT-4o (2024-11-20)`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4o`,description:`The 2024-11-20 version of GPT-4o offers a leveled-up creative writing ability with more natural, engaging, and tailored writing to improve relevance & readability.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0},pricing:{input:2.5,output:10,cached_input:1.25},model_type:`chat`,release_date:`2024-11-20`,alias:`openai/gpt-4o`,tagline:`The 2024-11-20 version of GPT-4o offers a leveled-up creative writing ability with more natural, engaging, and tailored writing to improve relevance & readability.`,tools:[`function_calling`],knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4o-audio-preview`,name:`OpenAI: GPT-4o Audio`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4o`,description:`The gpt-4o-audio-preview model adds support for audio inputs as prompts.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`audio`,`text`],output:[`text`,`audio`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:2.5,output:10},model_type:`chat`,release_date:`2025-08-15`,tagline:`The gpt-4o-audio-preview model adds support for audio inputs as prompts.`,tools:[`function_calling`],knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4o:extended`,name:`OpenAI: GPT-4o (extended)`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4o`,description:`GPT-4o ("o" for "omni") is OpenAI's latest AI model, supporting both text and image inputs with text outputs.`,context_window:128e3,max_output_tokens:64e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:6,output:18},model_type:`chat`,release_date:`2024-05-13`,tagline:`GPT-4o ("o" for "omni") is OpenAI's latest AI model, supporting both text and image inputs with text outputs.`,tools:[`function_calling`]},{id:`openai/gpt-4o-mini-2024-07-18`,name:`OpenAI: GPT-4o-mini (2024-07-18)`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4o`,description:`GPT-4o mini is OpenAI's newest model after [GPT-4 Omni](/models/openai/gpt-4o), supporting both text and image inputs with text outputs.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0,fine_tuning:!0},pricing:{input:.15,output:.6,cached_input:.075},model_type:`chat`,release_date:`2024-07-18`,alias:`openai/gpt-4o-mini`,tagline:`GPT-4o mini is OpenAI's newest model after [GPT-4 Omni](/models/openai/gpt-4o), supporting both text and image inputs with text outputs.`,tools:[`function_calling`],knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4o-mini-search-preview`,name:`OpenAI: GPT-4o-mini Search Preview`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4o`,description:`GPT-4o mini Search Preview is a specialized model for web search in Chat Completions.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,streaming:!0,vision:!0},pricing:{input:.15,output:.6},model_type:`chat`,release_date:`2025-03-12`,tagline:`GPT-4o mini Search Preview is a specialized model for web search in Chat Completions.`,knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4o-mini`,name:`OpenAI: GPT-4o-mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4o`,description:`GPT-4o mini is OpenAI's newest model after [GPT-4 Omni](/models/openai/gpt-4o), supporting both text and image inputs with text outputs.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0,fine_tuning:!0},pricing:{input:.15,output:.6,cached_input:.075},model_type:`chat`,release_date:`2024-07-18`,tagline:`GPT-4o mini is OpenAI's newest model after [GPT-4 Omni](/models/openai/gpt-4o), supporting both text and image inputs with text outputs.`,tools:[`function_calling`],snapshots:[`openai/gpt-4o-mini-2024-07-18`],knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4o-search-preview`,name:`OpenAI: GPT-4o Search Preview`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4o`,description:`GPT-4o Search Previewis a specialized model for web search in Chat Completions.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,streaming:!0,vision:!0},pricing:{input:2.5,output:10},model_type:`chat`,release_date:`2025-03-12`,tagline:`GPT-4o Search Previewis a specialized model for web search in Chat Completions.`,knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-4o`,name:`OpenAI: GPT-4o`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-4o`,description:`GPT-4o ("o" for "omni") is OpenAI's latest AI model, supporting both text and image inputs with text outputs.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0,fine_tuning:!0},pricing:{input:2.5,output:10,cached_input:1.25},model_type:`chat`,release_date:`2024-05-13`,tagline:`The 2024-11-20 version of GPT-4o offers a leveled-up creative writing ability with more natural, engaging, and tailored writing to improve relevance & readability.`,tools:[`function_calling`],snapshots:[`openai/gpt-4o-2024-11-20`,`openai/gpt-4o-2024-08-06`,`openai/gpt-4o-2024-05-13`],knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5-chat`,name:`OpenAI: GPT-5 Chat`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5`,description:`GPT-5 Chat is designed for advanced, natural, multimodal, and context-aware conversations for enterprise applications.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:1.25,output:10,cached_input:.125},model_type:`chat`,release_date:`2025-08-07`,tagline:`GPT-5 Chat is designed for advanced, natural, multimodal, and context-aware conversations for enterprise applications.`},{id:`openai/gpt-5-codex`,name:`OpenAI: GPT-5 Codex`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5`,description:`GPT-5-Codex is a specialized version of GPT-5 optimized for software engineering and coding workflows.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:1.25,output:10,cached_input:.125},model_type:`chat`,release_date:`2025-09-23`,tagline:`GPT-5-Codex is a specialized version of GPT-5 optimized for software engineering and coding workflows.`,tools:[`function_calling`],max_input_tokens:272e3,knowledge_cutoff:`2024-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5-image-mini`,name:`OpenAI: GPT-5 Image Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5`,description:`GPT-5 Image Mini combines OpenAI's advanced language capabilities, powered by [GPT-5 Mini](https://openrouter.ai/openai/gpt-5-mini), with GPT Image 1 Mini for efficient image generation.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`image`,`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:2.5,output:2,cached_input:.25},model_type:`chat`,release_date:`2025-10-16`,tagline:`GPT-5 Image Mini combines OpenAI's advanced language capabilities, powered by [GPT-5 Mini](https://openrouter.ai/openai/gpt-5-mini), with GPT Image 1 Mini for efficient image generation.`,tools:[`function_calling`]},{id:`openai/gpt-5-image`,name:`OpenAI: GPT-5 Image`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5`,description:`[GPT-5](https://openrouter.ai/openai/gpt-5) Image combines OpenAI's GPT-5 model with state-of-the-art image generation capabilities.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`image`,`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:10,output:10,cached_input:1.25},model_type:`chat`,release_date:`2025-10-14`,tagline:`[GPT-5](https://openrouter.ai/openai/gpt-5) Image combines OpenAI's GPT-5 model with state-of-the-art image generation capabilities.`,tools:[`function_calling`]},{id:`openai/gpt-5-mini`,name:`OpenAI: GPT-5 Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5`,description:`GPT-5 Mini is a compact version of GPT-5, designed to handle lighter-weight reasoning tasks.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:.25,output:2,cached_input:.025},model_type:`chat`,release_date:`2025-08-07`,tagline:`GPT-5 Mini is a compact version of GPT-5, designed to handle lighter-weight reasoning tasks.`,tools:[`function_calling`],max_input_tokens:272e3,knowledge_cutoff:`2024-05`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5-nano`,name:`OpenAI: GPT-5 Nano`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5`,description:`GPT-5-Nano is the smallest and fastest variant in the GPT-5 system, optimized for developer tools, rapid interactions, and ultra-low latency environments.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:.05,output:.4,cached_input:.005},model_type:`chat`,release_date:`2025-08-07`,tagline:`GPT-5-Nano is the smallest and fastest variant in the GPT-5 system, optimized for developer tools, rapid interactions, and ultra-low latency environments.`,tools:[`function_calling`],max_input_tokens:272e3,knowledge_cutoff:`2024-05`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5-pro`,name:`OpenAI: GPT-5 Pro`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5`,description:`GPT-5 Pro is OpenAI’s most advanced model, offering major improvements in reasoning, code quality, and user experience.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:15,output:120},model_type:`chat`,release_date:`2025-10-06`,tagline:`GPT-5 Pro is OpenAI’s most advanced model, offering major improvements in reasoning, code quality, and user experience.`,tools:[`function_calling`],knowledge_cutoff:`2024-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5.1-chat`,name:`OpenAI: GPT-5.1 Chat`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.1`,description:`GPT-5.1 Chat (AKA Instant is the fast, lightweight member of the 5.1 family, optimized for low-latency chat while retaining strong general intelligence.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:1.25,output:10,cached_input:.125},model_type:`chat`,release_date:`2025-11-13`,tagline:`GPT-5.1 Chat (AKA Instant is the fast, lightweight member of the 5.1 family, optimized for low-latency chat while retaining strong general intelligence.`,tools:[`function_calling`]},{id:`openai/gpt-5.1-codex-max`,name:`OpenAI: GPT-5.1-Codex-Max`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.1`,description:`GPT-5.1-Codex-Max is OpenAI’s latest agentic coding model, designed for long-running, high-context software development tasks.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:1.25,output:10,cached_input:.125},model_type:`chat`,release_date:`2025-12-04`,tagline:`GPT-5.1-Codex-Max is OpenAI’s latest agentic coding model, designed for long-running, high-context software development tasks.`,tools:[`function_calling`],knowledge_cutoff:`2024-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5.1-codex-mini`,name:`OpenAI: GPT-5.1-Codex-Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.1`,description:`GPT-5.1-Codex-Mini is a smaller and faster version of GPT-5.1-Codex`,context_window:4e5,max_output_tokens:1e5,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:.25,output:2,cached_input:.025},model_type:`chat`,release_date:`2025-11-13`,tagline:`GPT-5.1-Codex-Mini is a smaller and faster version of GPT-5.1-Codex`,tools:[`function_calling`],knowledge_cutoff:`2024-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5.1-codex`,name:`OpenAI: GPT-5.1-Codex`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.1`,description:`GPT-5.1-Codex is a specialized version of GPT-5.1 optimized for software engineering and coding workflows.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:1.25,output:10,cached_input:.125},model_type:`chat`,release_date:`2025-11-13`,tagline:`GPT-5.1-Codex is a specialized version of GPT-5.1 optimized for software engineering and coding workflows.`,tools:[`function_calling`],knowledge_cutoff:`2024-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5.1`,name:`OpenAI: GPT-5.1`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.1`,description:`GPT-5.1 is the latest frontier-grade model in the GPT-5 series, offering stronger general-purpose reasoning, improved instruction adherence, and a more natural conversational style compared to GPT-5.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:1.25,output:10,cached_input:.125},model_type:`chat`,release_date:`2025-11-13`,tagline:`GPT-5.1 is the latest frontier-grade model in the GPT-5 series, offering stronger general-purpose reasoning, improved instruction adherence, and a more natural conversational style compared to GPT-5.`,tools:[`function_calling`],knowledge_cutoff:`2024-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5.2-chat`,name:`OpenAI: GPT-5.2 Chat`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.2`,description:`GPT-5.2 Chat (AKA Instant) is the fast, lightweight member of the 5.2 family, optimized for low-latency chat while retaining strong general intelligence.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:1.75,output:14,cached_input:.175},model_type:`chat`,release_date:`2025-12-10`,tagline:`GPT-5.2 Chat (AKA Instant) is the fast, lightweight member of the 5.2 family, optimized for low-latency chat while retaining strong general intelligence.`,tools:[`function_calling`]},{id:`openai/gpt-5.2-codex`,name:`OpenAI: GPT-5.2-Codex`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.2`,description:`GPT-5.2-Codex is an upgraded version of GPT-5.1-Codex optimized for software engineering and coding workflows.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:1.75,output:14,cached_input:.175},model_type:`chat`,release_date:`2026-01-14`,tagline:`GPT-5.2-Codex is an upgraded version of GPT-5.1-Codex optimized for software engineering and coding workflows.`,tools:[`function_calling`],max_input_tokens:272e3,knowledge_cutoff:`2025-08`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5.2-pro`,name:`OpenAI: GPT-5.2 Pro`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.2`,description:`GPT-5.2 Pro is OpenAI’s most advanced model, offering major improvements in agentic coding and long context performance over GPT-5 Pro.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:21,output:168},model_type:`chat`,release_date:`2025-12-10`,tagline:`GPT-5.2 Pro is OpenAI’s most advanced model, offering major improvements in agentic coding and long context performance over GPT-5 Pro.`,tools:[`function_calling`],knowledge_cutoff:`2025-08`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5.2`,name:`OpenAI: GPT-5.2`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.2`,description:`GPT-5.2 is the latest frontier-grade model in the GPT-5 series, offering stronger agentic and long context perfomance compared to GPT-5.1.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:1.75,output:14,cached_input:.175},model_type:`chat`,release_date:`2025-12-10`,tagline:`GPT-5.2 is the latest frontier-grade model in the GPT-5 series, offering stronger agentic and long context perfomance compared to GPT-5.1.`,tools:[`function_calling`],knowledge_cutoff:`2025-08`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5.3-chat`,name:`OpenAI: GPT-5.3 Chat`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.3`,description:`GPT-5.3 Chat is an update to ChatGPT's most-used model that makes everyday conversations smoother, more useful, and more directly helpful.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:1.75,output:14,cached_input:.175},model_type:`chat`,release_date:`2026-03-03`,tagline:`GPT-5.3 Chat is an update to ChatGPT's most-used model that makes everyday conversations smoother, more useful, and more directly helpful.`,tools:[`function_calling`]},{id:`openai/gpt-5.3-codex`,name:`OpenAI: GPT-5.3-Codex`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.3`,description:`GPT-5.3-Codex is OpenAI’s most advanced agentic coding model, combining the frontier software engineering performance of GPT-5.2-Codex with the broader reasoning and professional knowledge capabilities of GPT-5.2.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:1.75,output:14,cached_input:.175},model_type:`chat`,release_date:`2026-02-24`,tagline:`GPT-5.3-Codex is OpenAI’s most advanced agentic coding model, combining the frontier software engineering performance of GPT-5.2-Codex with the broader reasoning and professional knowledge capabilitie`,tools:[`function_calling`],max_input_tokens:272e3,knowledge_cutoff:`2025-08`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5.4-mini`,name:`OpenAI: GPT-5.4 Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.4`,description:`GPT-5.4 mini brings the core capabilities of GPT-5.4 to a faster, more efficient model optimized for high-throughput workloads.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:.75,output:4.5,cached_input:.075},model_type:`chat`,release_date:`2026-03-17`,tagline:`GPT-5.4 mini brings the core capabilities of GPT-5.4 to a faster, more efficient model optimized for high-throughput workloads.`,tools:[`function_calling`],max_input_tokens:272e3,knowledge_cutoff:`2025-08`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5.4-nano`,name:`OpenAI: GPT-5.4 Nano`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.4`,description:`GPT-5.4 nano is the most lightweight and cost-efficient variant of the GPT-5.4 family, optimized for speed-critical and high-volume tasks.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:.2,output:1.25,cached_input:.02},model_type:`chat`,release_date:`2026-03-17`,tagline:`GPT-5.4 nano is the most lightweight and cost-efficient variant of the GPT-5.4 family, optimized for speed-critical and high-volume tasks.`,tools:[`function_calling`],max_input_tokens:272e3,knowledge_cutoff:`2025-08`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5.4-pro`,name:`OpenAI: GPT-5.4 Pro`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.4`,description:`GPT-5.4 Pro is OpenAI's most advanced model, building on GPT-5.4's unified architecture with enhanced reasoning capabilities for complex, high-stakes tasks.`,context_window:105e4,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:30,output:180},model_type:`chat`,release_date:`2026-03-05`,tagline:`GPT-5.4 Pro is OpenAI's most advanced model, building on GPT-5.4's unified architecture with enhanced reasoning capabilities for complex, high-stakes tasks.`,tools:[`function_calling`],knowledge_cutoff:`2025-08`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5.4`,name:`OpenAI: GPT-5.4`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5.4`,description:`GPT-5.4 is OpenAI’s latest frontier model, unifying the Codex and GPT lines into a single system.`,context_window:105e4,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:2.5,output:15,cached_input:.25},model_type:`chat`,release_date:`2026-03-05`,tagline:`GPT-5.4 is OpenAI’s latest frontier model, unifying the Codex and GPT lines into a single system.`,tools:[`function_calling`],knowledge_cutoff:`2025-08`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-5`,name:`OpenAI: GPT-5`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-5`,description:`GPT-5 is OpenAI’s most advanced model, offering major improvements in reasoning, code quality, and user experience.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:1.25,output:10,cached_input:.125},model_type:`chat`,release_date:`2025-08-07`,tagline:`GPT-5 is OpenAI’s most advanced model, offering major improvements in reasoning, code quality, and user experience.`,tools:[`function_calling`],max_input_tokens:272e3,knowledge_cutoff:`2024-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-audio-mini`,name:`OpenAI: GPT Audio Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-audio`,description:`A cost-efficient version of GPT Audio.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`audio`],output:[`text`,`audio`]},capabilities:{structured_output:!0,tool_call:!0},pricing:{input:.6,output:2.4},release_date:`2026-01-19`,tagline:`A cost-efficient version of GPT Audio.`,knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-audio`,name:`OpenAI: GPT Audio`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-audio`,description:`The gpt-audio model is OpenAI's first generally available audio model.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`audio`],output:[`text`,`audio`]},capabilities:{structured_output:!0,streaming:!0,tool_call:!0},pricing:{input:2.5,output:10},release_date:`2026-01-19`,tagline:`The gpt-audio model is OpenAI's first generally available audio model.`,knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/gpt-oss-120b:free`,name:`OpenAI: gpt-oss-120b (free)`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-oss`,description:`gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},model_type:`chat`,release_date:`2025-08-05`,tagline:`gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases.`,tools:[`function_calling`]},{id:`openai/gpt-oss-120b`,name:`OpenAI: gpt-oss-120b`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-oss`,description:`gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.039,output:.19},model_type:`chat`,parameters:120,release_date:`2025-08-05`,tagline:`gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases.`,tools:[`function_calling`],max_output_tokens:131072,knowledge_cutoff:`2024-05`,license:`apache-2.0`,open_weight:!0},{id:`openai/gpt-oss-20b:free`,name:`OpenAI: gpt-oss-20b (free)`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-oss`,description:`gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},model_type:`chat`,release_date:`2025-08-05`,tagline:`gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license.`,tools:[`function_calling`]},{id:`openai/gpt-oss-20b`,name:`OpenAI: gpt-oss-20b`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-oss`,description:`gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.03,output:.11,cached_input:.015},model_type:`chat`,parameters:20,release_date:`2025-08-05`,tagline:`gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license.`,tools:[`function_calling`],knowledge_cutoff:`2024-05`,license:`apache-2.0`,open_weight:!0},{id:`openai/gpt-oss-safeguard-20b`,name:`OpenAI: gpt-oss-safeguard-20b`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`gpt-oss`,description:`gpt-oss-safeguard-20b is a safety reasoning model from OpenAI built upon gpt-oss-20b.`,context_window:131072,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.075,output:.3,cached_input:.037},model_type:`moderation`,parameters:20,release_date:`2025-10-29`,tagline:`gpt-oss-safeguard-20b is a safety reasoning model from OpenAI built upon gpt-oss-20b.`,tools:[`function_calling`]},{id:`openai/o1-pro`,name:`OpenAI: o1-pro`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`o1`,description:`The o1 series of models are trained with reinforcement learning to think before they answer and perform complex reasoning.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0,vision:!0,tool_call:!0},pricing:{input:150,output:600},model_type:`reasoning`,release_date:`2025-03-19`,tagline:`The o1 series of models are trained with reinforcement learning to think before they answer and perform complex reasoning.`,knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/o1`,name:`OpenAI: o1`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`o1`,description:`The latest and strongest model family from OpenAI, o1 is designed to spend more time thinking before responding.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:15,output:60,cached_input:7.5},model_type:`reasoning`,release_date:`2024-12-17`,tagline:`The latest and strongest model family from OpenAI, o1 is designed to spend more time thinking before responding.`,tools:[`function_calling`],knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/o3-deep-research`,name:`OpenAI: o3 Deep Research`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`o3`,description:`o3-deep-research is OpenAI's advanced model for deep research, designed to tackle complex, multi-step research tasks.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:10,output:40,cached_input:2.5},model_type:`chat`,release_date:`2025-10-10`,tagline:`o3-deep-research is OpenAI's advanced model for deep research, designed to tackle complex, multi-step research tasks.`,tools:[`function_calling`],knowledge_cutoff:`2024-05`,license:`proprietary`,open_weight:!1},{id:`openai/o3-mini-high`,name:`OpenAI: o3 Mini High`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`o3`,description:`OpenAI o3-mini-high is the same model as [o3-mini](/openai/o3-mini) with reasoning_effort set to high.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.1,output:4.4,cached_input:.55},model_type:`reasoning`,release_date:`2025-02-12`,tagline:`OpenAI o3-mini-high is the same model as [o3-mini](/openai/o3-mini) with reasoning_effort set to high.`,tools:[`function_calling`]},{id:`openai/o3-mini`,name:`OpenAI: o3 Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`o3`,description:`OpenAI o3-mini is a cost-efficient language model optimized for STEM reasoning tasks, particularly excelling in science, mathematics, and coding.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:1.1,output:4.4,cached_input:.55},model_type:`reasoning`,release_date:`2025-01-31`,tagline:`OpenAI o3-mini is a cost-efficient language model optimized for STEM reasoning tasks, particularly excelling in science, mathematics, and coding.`,tools:[`function_calling`],knowledge_cutoff:`2023-09`,license:`proprietary`,open_weight:!1},{id:`openai/o3-pro`,name:`OpenAI: o3 Pro`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`o3`,description:`The o-series of models are trained with reinforcement learning to think before they answer and perform complex reasoning.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,vision:!0},pricing:{input:20,output:80},model_type:`reasoning`,release_date:`2025-06-10`,tagline:`The o-series of models are trained with reinforcement learning to think before they answer and perform complex reasoning.`,tools:[`function_calling`],knowledge_cutoff:`2024-05`,license:`proprietary`,open_weight:!1},{id:`openai/o3`,name:`OpenAI: o3`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`o3`,description:`o3 is a well-rounded and powerful model across domains.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:2,output:8,cached_input:.5},model_type:`reasoning`,release_date:`2025-04-16`,tagline:`o3 is a well-rounded and powerful model across domains.`,tools:[`function_calling`],knowledge_cutoff:`2024-05`,license:`proprietary`,open_weight:!1},{id:`openai/o4-mini-deep-research`,name:`OpenAI: o4 Mini Deep Research`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`o4`,description:`o4-mini-deep-research is OpenAI's faster, more affordable deep research model—ideal for tackling complex, multi-step research tasks.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:2,output:8,cached_input:.5},model_type:`chat`,release_date:`2025-10-10`,tagline:`o4-mini-deep-research is OpenAI's faster, more affordable deep research model—ideal for tackling complex, multi-step research tasks.`,tools:[`function_calling`],knowledge_cutoff:`2024-05`,license:`proprietary`,open_weight:!1},{id:`openai/o4-mini-high`,name:`OpenAI: o4 Mini High`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`o4`,description:`OpenAI o4-mini-high is the same model as [o4-mini](/openai/o4-mini) with reasoning_effort set to high.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.1,output:4.4,cached_input:.275},model_type:`reasoning`,release_date:`2025-04-16`,tagline:`OpenAI o4-mini-high is the same model as [o4-mini](/openai/o4-mini) with reasoning_effort set to high.`,tools:[`function_calling`]},{id:`openai/o4-mini`,name:`OpenAI: o4 Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-23`,family:`o4`,description:`OpenAI o4-mini is a compact reasoning model in the o-series, optimized for fast, cost-efficient performance while retaining strong multimodal and agentic capabilities.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0,fine_tuning:!0},pricing:{input:1.1,output:4.4,cached_input:.275},model_type:`reasoning`,release_date:`2025-04-16`,tagline:`OpenAI o4-mini is a compact reasoning model in the o-series, optimized for fast, cost-efficient performance while retaining strong multimodal and agentic capabilities.`,tools:[`function_calling`],knowledge_cutoff:`2024-05`,license:`proprietary`,open_weight:!1},{id:`openrouter/auto`,name:`Auto Router`,created_by:`openrouter`,source:`official`,last_updated:`2026-03-23`,description:`Your prompt will be processed by a meta-model and routed to one of dozens of models (see below), optimizing for the best possible output.`,context_window:2e6,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`,`image`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},release_date:`2023-11-08`,tagline:`Your prompt will be processed by a meta-model and routed to one of dozens of models (see below), optimizing for the best possible output.`,tools:[`function_calling`]},{id:`openrouter/bodybuilder`,name:`Body Builder (beta)`,created_by:`openrouter`,source:`official`,last_updated:`2026-03-23`,description:`Transform your natural language requests into structured OpenRouter API request objects.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]},release_date:`2025-12-05`,tagline:`Transform your natural language requests into structured OpenRouter API request objects.`},{id:`openrouter/free`,name:`Free Models Router`,created_by:`openrouter`,source:`official`,last_updated:`2026-03-23`,description:`The simplest way to get free inference.`,context_window:2e5,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},release_date:`2026-02-01`,tagline:`The simplest way to get free inference.`,tools:[`function_calling`]},{id:`perplexity/sonar-deep-research`,name:`Perplexity: Sonar Deep Research`,created_by:`perplexity`,source:`official`,last_updated:`2026-03-23`,description:`Sonar Deep Research is a research-focused model designed for multi-step retrieval, synthesis, and reasoning across complex topics.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0,streaming:!0,tool_call:!0},pricing:{input:2,output:8},family:`sonar`,model_type:`chat`,release_date:`2025-03-07`,tagline:`Sonar Deep Research is a research-focused model designed for multi-step retrieval, synthesis, and reasoning across complex topics.`,license:`proprietary`,open_weight:!1},{id:`perplexity/sonar-pro-search`,name:`Perplexity: Sonar Pro Search`,created_by:`perplexity`,source:`official`,last_updated:`2026-03-23`,description:`Exclusively available on the OpenRouter API, Sonar Pro's new Pro Search mode is Perplexity's most advanced agentic search system.`,context_window:2e5,max_output_tokens:8e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:3,output:15},family:`sonar`,model_type:`chat`,release_date:`2025-10-30`,tagline:`Exclusively available on the OpenRouter API, Sonar Pro's new Pro Search mode is Perplexity's most advanced agentic search system.`},{id:`perplexity/sonar-pro`,name:`Perplexity: Sonar Pro`,created_by:`perplexity`,source:`official`,last_updated:`2026-03-23`,description:`Note: Sonar Pro pricing includes Perplexity search pricing.`,context_window:2e5,max_output_tokens:8e3,modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:3,output:15},family:`sonar`,model_type:`chat`,release_date:`2025-03-07`,tagline:`Note: Sonar Pro pricing includes Perplexity search pricing.`,license:`proprietary`,open_weight:!1,capabilities:{streaming:!0,tool_call:!0}},{id:`perplexity/sonar-reasoning-pro`,name:`Perplexity: Sonar Reasoning Pro`,created_by:`perplexity`,source:`official`,last_updated:`2026-03-23`,description:`Note: Sonar Pro pricing includes Perplexity search pricing.`,context_window:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{reasoning:!0,streaming:!0,tool_call:!0},pricing:{input:2,output:8},model_type:`reasoning`,family:`sonar`,release_date:`2025-03-07`,tagline:`Note: Sonar Pro pricing includes Perplexity search pricing.`,license:`proprietary`,open_weight:!1},{id:`perplexity/sonar`,name:`Perplexity: Sonar`,created_by:`perplexity`,source:`official`,last_updated:`2026-03-23`,description:`Sonar is lightweight, affordable, fast, and simple to use — now featuring citations and the ability to customize sources.`,context_window:127072,modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:1,output:1},family:`sonar`,model_type:`chat`,release_date:`2025-01-27`,tagline:`Sonar is lightweight, affordable, fast, and simple to use — now featuring citations and the ability to customize sources.`,license:`proprietary`,open_weight:!1,capabilities:{streaming:!0,tool_call:!0}},{id:`prime-intellect/intellect-3`,name:`Prime Intellect: INTELLECT-3`,created_by:`prime-intellect`,source:`official`,last_updated:`2026-03-23`,description:`INTELLECT-3 is a 106B-parameter Mixture-of-Experts model (12B active) post-trained from GLM-4.5-Air-Base using supervised fine-tuning (SFT) followed by large-scale reinforcement learning (RL).`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.2,output:1.1},release_date:`2025-11-27`,tagline:`INTELLECT-3 is a 106B-parameter Mixture-of-Experts model (12B active) post-trained from GLM-4.5-Air-Base using supervised fine-tuning (SFT) followed by large-scale reinforcement learning (RL).`,tools:[`function_calling`]},{id:`qwen/qwen-2.5-72b-instruct`,name:`Qwen2.5 72B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen`,description:`Qwen2.5 72B is the latest series of Qwen large language models.`,context_window:32768,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:.12,output:.39},model_type:`chat`,parameters:72,release_date:`2024-09-19`,tagline:`Qwen2.5 72B is the latest series of Qwen large language models.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen-2.5-7b-instruct`,name:`Qwen: Qwen2.5 7B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen`,description:`Qwen2.5 7B is the latest series of Qwen large language models.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:.04,output:.1},model_type:`chat`,parameters:7,release_date:`2024-10-16`,tagline:`Qwen2.5 7B is the latest series of Qwen large language models.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen-2.5-coder-32b-instruct`,name:`Qwen2.5 Coder 32B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen`,description:`Qwen2.5-Coder is the latest series of Code-Specific Qwen large language models (formerly known as CodeQwen).`,context_window:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.66,output:1},model_type:`code`,parameters:32,release_date:`2024-11-11`,tagline:`Qwen2.5-Coder is the latest series of Code-Specific Qwen large language models (formerly known as CodeQwen).`},{id:`qwen/qwen-2.5-vl-7b-instruct`,name:`Qwen: Qwen2.5-VL 7B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen`,description:`Qwen2.5 VL 7B is a multimodal LLM from the Qwen Team with the following key enhancements:
+
+- SoTA understanding of images of various resolution & ratio: Qwen2.5-VL achieves state-of-the-art performanc`,context_window:32768,modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:.2,output:.2},model_type:`chat`,parameters:7,release_date:`2024-08-28`,tagline:`Qwen2.5 VL 7B is a multimodal LLM from the Qwen Team with the following key enhancements:
 
-It provides a significant upgrade on the pr`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2,output:6,cached_input:.2}},{id:`mistralai/mistral-large-2512`,name:`Mistral: Mistral Large 3 2512`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mistral-large`,description:`Mistral Large 3 2512 is Mistral’s most capable model to date, featuring a sparse mixture-of-experts architecture with 41B active parameters (675B total), and released under the Apache 2.0 license.`,context_window:262144,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.5,output:1.5,cached_input:.05}},{id:`mistralai/mistral-large`,name:`Mistral Large`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mistral-large`,description:"This is Mistral AI's flagship model, Mistral Large 2 (version `mistral-large-2407`).",context_window:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2,output:6,cached_input:.2}},{id:`mistralai/mistral-medium-3.1`,name:`Mistral: Mistral Medium 3.1`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mistral-medium`,description:`Mistral Medium 3.1 is an updated version of Mistral Medium 3, which is a high-performance enterprise-grade language model designed to deliver frontier-level capabilities at significantly reduced operational cost.`,context_window:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:2,cached_input:.04}},{id:`mistralai/mistral-medium-3`,name:`Mistral: Mistral Medium 3`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mistral-medium`,description:`Mistral Medium 3 is a high-performance enterprise-grade language model designed to deliver frontier-level capabilities at significantly reduced operational cost.`,context_window:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:2,cached_input:.04}},{id:`mistralai/mistral-nemo`,name:`Mistral: Mistral Nemo`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,description:`A 12B parameter model with a 128k token context length built by Mistral in collaboration with NVIDIA.`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.02,output:.04}},{id:`mistralai/mistral-saba`,name:`Mistral: Saba`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,description:`Mistral Saba is a 24B-parameter language model specifically designed for the Middle East and South Asia, delivering accurate and contextually relevant responses while maintaining efficient performance.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.2,output:.6,cached_input:.02}},{id:`mistralai/mistral-small-24b-instruct-2501`,name:`Mistral: Mistral Small 3`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mistral-small`,description:`Mistral Small 3 is a 24B-parameter language model optimized for low-latency performance across common AI tasks.`,context_window:32768,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.05,output:.08}},{id:`mistralai/mistral-small-2603`,name:`Mistral: Mistral Small 4`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mistral-small`,description:`Mistral Small 4 is the next major release in the Mistral Small family, unifying the capabilities of several flagship Mistral models into a single system.`,context_window:262144,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.15,output:.6,cached_input:.015}},{id:`mistralai/mistral-small-3.1-24b-instruct:free`,name:`Mistral: Mistral Small 3.1 24B (free)`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mistral-small`,description:`Mistral Small 3.1 24B Instruct is an upgraded variant of Mistral Small 3 (2501), featuring 24 billion parameters with advanced multimodal capabilities.`,status:`deprecated`,deprecation_date:`2026-03-29`,context_window:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0}},{id:`mistralai/mistral-small-3.1-24b-instruct`,name:`Mistral: Mistral Small 3.1 24B`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mistral-small`,description:`Mistral Small 3.1 24B Instruct is an upgraded variant of Mistral Small 3 (2501), featuring 24 billion parameters with advanced multimodal capabilities.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.03,output:.11,cached_input:.015}},{id:`mistralai/mistral-small-3.2-24b-instruct`,name:`Mistral: Mistral Small 3.2 24B`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mistral-small`,description:`Mistral-Small-3.2-24B-Instruct-2506 is an updated 24B parameter model from Mistral optimized for instruction following, repetition reduction, and improved function calling.`,context_window:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.075,output:.2}},{id:`mistralai/mistral-small-creative`,name:`Mistral: Mistral Small Creative`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mistral-small`,description:`Mistral Small Creative is an experimental small model designed for creative writing, narrative generation, roleplay and character-driven dialogue, general-purpose instruction following, and conversati`,status:`deprecated`,deprecation_date:`2026-04-30`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.1,output:.3,cached_input:.01}},{id:`mistralai/mixtral-8x22b-instruct`,name:`Mistral: Mixtral 8x22B Instruct`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mixtral`,description:`Mistral's official instruct fine-tuned version of [Mixtral 8x22B](/models/mistralai/mixtral-8x22b).`,context_window:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2,output:6,cached_input:.2}},{id:`mistralai/mixtral-8x7b-instruct`,name:`Mistral: Mixtral 8x7B Instruct`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`mixtral`,description:`Mixtral 8x7B Instruct is a pretrained generative Sparse Mixture of Experts, by Mistral AI, for chat and instruction use.`,context_window:32768,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.54,output:.54}},{id:`mistralai/pixtral-12b`,name:`Mistral: Pixtral 12B`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`pixtral`,description:`The first multi-modal, text+image-to-text model from Mistral AI.`,context_window:32768,modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:.1,output:.1}},{id:`mistralai/pixtral-large-2411`,name:`Mistral: Pixtral Large 2411`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,family:`pixtral`,description:`Pixtral Large is a 124B parameter, open-weight, multimodal model built on top of [Mistral Large 2](/mistralai/mistral-large-2411).`,context_window:131072,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2,output:6,cached_input:.2}},{id:`mistralai/voxtral-small-24b-2507`,name:`Mistral: Voxtral Small 24B 2507`,created_by:`mistral`,source:`official`,last_updated:`2026-03-21`,description:`Voxtral Small is an enhancement of Mistral Small 3, incorporating state-of-the-art audio input capabilities while retaining best-in-class text performance.`,context_window:32e3,modalities:{input:[`text`,`audio`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.1,output:.3,cached_input:.01}},{id:`moonshotai/kimi-k2-0905`,name:`MoonshotAI: Kimi K2 0905`,created_by:`moonshotai`,source:`official`,last_updated:`2026-03-21`,description:`Kimi K2 0905 is the September update of [Kimi K2 0711](moonshotai/kimi-k2).`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:2,cached_input:.15}},{id:`moonshotai/kimi-k2-thinking`,name:`MoonshotAI: Kimi K2 Thinking`,created_by:`moonshotai`,source:`official`,last_updated:`2026-03-21`,description:`Kimi K2 Thinking is Moonshot AI’s most advanced open reasoning model to date, extending the K2 series into agentic, long-horizon reasoning.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.47,output:2,cached_input:.141}},{id:`moonshotai/kimi-k2.5`,name:`MoonshotAI: Kimi K2.5`,created_by:`moonshotai`,source:`official`,last_updated:`2026-03-21`,description:`Kimi K2.5 is Moonshot AI's native multimodal model, delivering state-of-the-art visual coding capability and a self-directed agent swarm paradigm.`,context_window:262144,max_output_tokens:65535,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.45,output:2.2,cached_input:.225}},{id:`moonshotai/kimi-k2`,name:`MoonshotAI: Kimi K2 0711`,created_by:`moonshotai`,source:`official`,last_updated:`2026-03-21`,description:`Kimi K2 Instruct is a large-scale Mixture-of-Experts (MoE) language model developed by Moonshot AI, featuring 1 trillion total parameters with 32 billion active per forward pass.`,context_window:131e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.55,output:2.2}},{id:`morph/morph-v3-fast`,name:`Morph: Morph V3 Fast`,created_by:`morph`,source:`official`,last_updated:`2026-03-21`,description:`Morph's fastest apply model for code edits.`,context_window:81920,max_output_tokens:38e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:.8,output:1.2}},{id:`morph/morph-v3-large`,name:`Morph: Morph V3 Large`,created_by:`morph`,source:`official`,last_updated:`2026-03-21`,description:`Morph's high-accuracy apply model for complex code edits.`,context_window:262144,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},pricing:{input:.9,output:1.9}},{id:`nex-agi/deepseek-v3.1-nex-n1`,name:`Nex AGI: DeepSeek V3.1 Nex N1`,created_by:`nex-agi`,source:`official`,last_updated:`2026-03-21`,family:`deepseek`,description:`DeepSeek V3.1 Nex-N1 is the flagship release of the Nex-N1 series — a post-trained model designed to highlight agent autonomy, tool use, and real-world productivity.`,context_window:131072,max_output_tokens:163840,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.27,output:1}},{id:`nousresearch/hermes-2-pro-llama-3-8b`,name:`NousResearch: Hermes 2 Pro - Llama-3 8B`,created_by:`nousresearch`,source:`official`,last_updated:`2026-03-21`,description:`Hermes 2 Pro is an upgraded, retrained version of Nous Hermes 2, consisting of an updated and cleaned version of the OpenHermes 2.5 Dataset, as well as a newly introduced Function Calling and JSON Mod`,context_window:8192,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.14,output:.14}},{id:`nousresearch/hermes-3-llama-3.1-405b:free`,name:`Nous: Hermes 3 405B Instruct (free)`,created_by:`nousresearch`,source:`official`,last_updated:`2026-03-21`,description:`Hermes 3 is a generalist language model with many improvements over Hermes 2, including advanced agentic capabilities, much better roleplaying, reasoning, multi-turn conversation, long context coherence, and improvements across the board.`,context_window:131072,modalities:{input:[`text`],output:[`text`]}},{id:`nousresearch/hermes-3-llama-3.1-405b`,name:`Nous: Hermes 3 405B Instruct`,created_by:`nousresearch`,source:`official`,last_updated:`2026-03-21`,description:`Hermes 3 is a generalist language model with many improvements over Hermes 2, including advanced agentic capabilities, much better roleplaying, reasoning, multi-turn conversation, long context coherence, and improvements across the board.`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:1,output:1}},{id:`nousresearch/hermes-3-llama-3.1-70b`,name:`Nous: Hermes 3 70B Instruct`,created_by:`nousresearch`,source:`official`,last_updated:`2026-03-21`,description:`Hermes 3 is a generalist language model with many improvements over [Hermes 2](/models/nousresearch/nous-hermes-2-mistral-7b-dpo), including advanced agentic capabilities, much better roleplaying, reasoning, multi-turn conversation, long context coherence, and improvements across the board.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.3,output:.3}},{id:`nousresearch/hermes-4-405b`,name:`Nous: Hermes 4 405B`,created_by:`nousresearch`,source:`official`,last_updated:`2026-03-21`,description:`Hermes 4 is a large-scale reasoning model built on Meta-Llama-3.1-405B and released by Nous Research.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:1,output:3}},{id:`nousresearch/hermes-4-70b`,name:`Nous: Hermes 4 70B`,created_by:`nousresearch`,source:`official`,last_updated:`2026-03-21`,description:`Hermes 4 70B is a hybrid reasoning model from Nous Research, built on Meta-Llama-3.1-70B.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.13,output:.4}},{id:`nvidia/llama-3.1-nemotron-70b-instruct`,name:`NVIDIA: Llama 3.1 Nemotron 70B Instruct`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-21`,family:`llama-3.1`,description:`NVIDIA's Llama 3.1 Nemotron 70B is a language model designed for generating precise and useful responses.`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:1.2,output:1.2}},{id:`nvidia/llama-3.1-nemotron-ultra-253b-v1`,name:`NVIDIA: Llama 3.1 Nemotron Ultra 253B v1`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-21`,family:`llama-3.1`,description:`Llama-3.1-Nemotron-Ultra-253B-v1 is a large language model (LLM) optimized for advanced reasoning, human-interactive chat, retrieval-augmented generation (RAG), and tool-calling tasks.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.6,output:1.8}},{id:`nvidia/llama-3.3-nemotron-super-49b-v1.5`,name:`NVIDIA: Llama 3.3 Nemotron Super 49B V1.5`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-21`,family:`llama-3.3`,description:`Llama-3.3-Nemotron-Super-49B-v1.5 is a 49B-parameter, English-centric reasoning/chat model derived from Meta’s Llama-3.3-70B-Instruct with a 128K context.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.1,output:.4}},{id:`nvidia/nemotron-3-nano-30b-a3b:free`,name:`NVIDIA: Nemotron 3 Nano 30B A3B (free)`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-21`,description:`NVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems.`,context_window:256e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0}},{id:`nvidia/nemotron-3-nano-30b-a3b`,name:`NVIDIA: Nemotron 3 Nano 30B A3B`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-21`,description:`NVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.05,output:.2}},{id:`nvidia/nemotron-3-super-120b-a12b:free`,name:`NVIDIA: Nemotron 3 Super (free)`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-21`,description:`NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications.`,context_window:262144,max_output_tokens:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0}},{id:`nvidia/nemotron-3-super-120b-a12b`,name:`NVIDIA: Nemotron 3 Super`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-21`,description:`NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.1,output:.5,cached_input:.04}},{id:`nvidia/nemotron-nano-12b-v2-vl:free`,name:`NVIDIA: Nemotron Nano 12B 2 VL (free)`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-21`,description:`NVIDIA Nemotron Nano 2 VL is a 12-billion-parameter open multimodal reasoning model designed for video understanding and document intelligence.`,context_window:128e3,max_output_tokens:128e3,modalities:{input:[`image`,`text`,`video`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0}},{id:`nvidia/nemotron-nano-12b-v2-vl`,name:`NVIDIA: Nemotron Nano 12B 2 VL`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-21`,description:`NVIDIA Nemotron Nano 2 VL is a 12-billion-parameter open multimodal reasoning model designed for video understanding and document intelligence.`,context_window:131072,modalities:{input:[`image`,`text`,`video`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.2,output:.6}},{id:`nvidia/nemotron-nano-9b-v2:free`,name:`NVIDIA: Nemotron Nano 9B V2 (free)`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-21`,description:`NVIDIA-Nemotron-Nano-9B-v2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0}},{id:`nvidia/nemotron-nano-9b-v2`,name:`NVIDIA: Nemotron Nano 9B V2`,created_by:`nvidia`,source:`official`,last_updated:`2026-03-21`,description:`NVIDIA-Nemotron-Nano-9B-v2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.04,output:.16}},{id:`openai/gpt-3.5-turbo-0613`,name:`OpenAI: GPT-3.5 Turbo (older v0613)`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-3.5`,description:`GPT-3.5 Turbo is OpenAI's fastest model.`,context_window:4095,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:1,output:2}},{id:`openai/gpt-3.5-turbo-16k`,name:`OpenAI: GPT-3.5 Turbo 16k`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-3.5`,description:`This model offers four times the context length of gpt-3.5-turbo, allowing it to support approximately 20 pages of text in a single request at a higher cost.`,context_window:16385,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:3,output:4}},{id:`openai/gpt-3.5-turbo-instruct`,name:`OpenAI: GPT-3.5 Turbo Instruct`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-3.5`,description:`This model is a variant of GPT-3.5 Turbo tuned for instructional prompts and omitting chat-related optimizations.`,context_window:4095,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:1.5,output:2}},{id:`openai/gpt-3.5-turbo`,name:`OpenAI: GPT-3.5 Turbo`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-3.5`,description:`GPT-3.5 Turbo is OpenAI's fastest model.`,context_window:16385,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.5,output:1.5}},{id:`openai/gpt-4-0314`,name:`OpenAI: GPT-4 (older v0314)`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4`,description:`GPT-4-0314 is the first version of GPT-4 released, with a context length of 8,192 tokens, and was supported until June 14.`,context_window:8191,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:30,output:60}},{id:`openai/gpt-4-1106-preview`,name:`OpenAI: GPT-4 Turbo (older v1106)`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4`,description:`The latest GPT-4 Turbo model with vision capabilities.`,context_window:128e3,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:10,output:30}},{id:`openai/gpt-4-turbo-preview`,name:`OpenAI: GPT-4 Turbo Preview`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4-turbo`,description:`The preview GPT-4 model with improved instruction following, JSON mode, reproducible outputs, parallel function calling, and more.`,context_window:128e3,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:10,output:30}},{id:`openai/gpt-4-turbo`,name:`OpenAI: GPT-4 Turbo`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4-turbo`,description:`The latest GPT-4 Turbo model with vision capabilities.`,context_window:128e3,max_output_tokens:4096,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:10,output:30}},{id:`openai/gpt-4.1-mini`,name:`OpenAI: GPT-4.1 Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4.1`,description:`GPT-4.1 Mini is a mid-sized model delivering performance competitive with GPT-4o at substantially lower latency and cost.`,context_window:1047576,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:1.6,cached_input:.1}},{id:`openai/gpt-4.1-nano`,name:`OpenAI: GPT-4.1 Nano`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4.1`,description:`For tasks that demand low latency, GPT‑4.1 nano is the fastest and cheapest model in the GPT-4.1 series.`,context_window:1047576,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.1,output:.4,cached_input:.025}},{id:`openai/gpt-4.1`,name:`OpenAI: GPT-4.1`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4.1`,description:`GPT-4.1 is a flagship large language model optimized for advanced instruction following, real-world software engineering, and long-context reasoning.`,context_window:1047576,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2,output:8,cached_input:.5}},{id:`openai/gpt-4`,name:`OpenAI: GPT-4`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4`,description:`OpenAI's flagship model, GPT-4 is a large-scale multimodal language model capable of solving difficult problems with greater accuracy than previous models due to its broader general knowledge and advanced reasoning capabilities.`,context_window:8191,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:30,output:60}},{id:`openai/gpt-4o-2024-05-13`,name:`OpenAI: GPT-4o (2024-05-13)`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4o`,description:`GPT-4o ("o" for "omni") is OpenAI's latest AI model, supporting both text and image inputs with text outputs.`,context_window:128e3,max_output_tokens:4096,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:5,output:15}},{id:`openai/gpt-4o-2024-08-06`,name:`OpenAI: GPT-4o (2024-08-06)`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4o`,description:`The 2024-08-06 version of GPT-4o offers improved performance in structured outputs, with the ability to supply a JSON schema in the respone_format.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2.5,output:10,cached_input:1.25}},{id:`openai/gpt-4o-2024-11-20`,name:`OpenAI: GPT-4o (2024-11-20)`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4o`,description:`The 2024-11-20 version of GPT-4o offers a leveled-up creative writing ability with more natural, engaging, and tailored writing to improve relevance & readability.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2.5,output:10,cached_input:1.25}},{id:`openai/gpt-4o-audio-preview`,name:`OpenAI: GPT-4o Audio`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4o`,description:`The gpt-4o-audio-preview model adds support for audio inputs as prompts.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`audio`,`text`],output:[`text`,`audio`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2.5,output:10}},{id:`openai/gpt-4o:extended`,name:`OpenAI: GPT-4o (extended)`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4o`,description:`GPT-4o ("o" for "omni") is OpenAI's latest AI model, supporting both text and image inputs with text outputs.`,context_window:128e3,max_output_tokens:64e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:6,output:18}},{id:`openai/gpt-4o-mini-2024-07-18`,name:`OpenAI: GPT-4o-mini (2024-07-18)`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4o`,description:`GPT-4o mini is OpenAI's newest model after [GPT-4 Omni](/models/openai/gpt-4o), supporting both text and image inputs with text outputs.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.15,output:.6,cached_input:.075}},{id:`openai/gpt-4o-mini-search-preview`,name:`OpenAI: GPT-4o-mini Search Preview`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4o`,description:`GPT-4o mini Search Preview is a specialized model for web search in Chat Completions.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.15,output:.6}},{id:`openai/gpt-4o-mini`,name:`OpenAI: GPT-4o-mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4o`,description:`GPT-4o mini is OpenAI's newest model after [GPT-4 Omni](/models/openai/gpt-4o), supporting both text and image inputs with text outputs.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.15,output:.6,cached_input:.075}},{id:`openai/gpt-4o-search-preview`,name:`OpenAI: GPT-4o Search Preview`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4o`,description:`GPT-4o Search Previewis a specialized model for web search in Chat Completions.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:2.5,output:10}},{id:`openai/gpt-4o`,name:`OpenAI: GPT-4o`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-4o`,description:`GPT-4o ("o" for "omni") is OpenAI's latest AI model, supporting both text and image inputs with text outputs.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:2.5,output:10,cached_input:1.25}},{id:`openai/gpt-5-chat`,name:`OpenAI: GPT-5 Chat`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5`,description:`GPT-5 Chat is designed for advanced, natural, multimodal, and context-aware conversations for enterprise applications.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:1.25,output:10,cached_input:.125}},{id:`openai/gpt-5-codex`,name:`OpenAI: GPT-5 Codex`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5`,description:`GPT-5-Codex is a specialized version of GPT-5 optimized for software engineering and coding workflows.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.25,output:10,cached_input:.125}},{id:`openai/gpt-5-image-mini`,name:`OpenAI: GPT-5 Image Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5`,description:`GPT-5 Image Mini combines OpenAI's advanced language capabilities, powered by [GPT-5 Mini](https://openrouter.ai/openai/gpt-5-mini), with GPT Image 1 Mini for efficient image generation.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`image`,`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:2.5,output:2,cached_input:.25}},{id:`openai/gpt-5-image`,name:`OpenAI: GPT-5 Image`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5`,description:`[GPT-5](https://openrouter.ai/openai/gpt-5) Image combines OpenAI's GPT-5 model with state-of-the-art image generation capabilities.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`image`,`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:10,output:10,cached_input:1.25}},{id:`openai/gpt-5-mini`,name:`OpenAI: GPT-5 Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5`,description:`GPT-5 Mini is a compact version of GPT-5, designed to handle lighter-weight reasoning tasks.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.25,output:2,cached_input:.025}},{id:`openai/gpt-5-nano`,name:`OpenAI: GPT-5 Nano`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5`,description:`GPT-5-Nano is the smallest and fastest variant in the GPT-5 system, optimized for developer tools, rapid interactions, and ultra-low latency environments.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.05,output:.4,cached_input:.005}},{id:`openai/gpt-5-pro`,name:`OpenAI: GPT-5 Pro`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5`,description:`GPT-5 Pro is OpenAI’s most advanced model, offering major improvements in reasoning, code quality, and user experience.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:15,output:120}},{id:`openai/gpt-5.1-chat`,name:`OpenAI: GPT-5.1 Chat`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.1`,description:`GPT-5.1 Chat (AKA Instant is the fast, lightweight member of the 5.1 family, optimized for low-latency chat while retaining strong general intelligence.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:1.25,output:10,cached_input:.125}},{id:`openai/gpt-5.1-codex-max`,name:`OpenAI: GPT-5.1-Codex-Max`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.1`,description:`GPT-5.1-Codex-Max is OpenAI’s latest agentic coding model, designed for long-running, high-context software development tasks.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.25,output:10,cached_input:.125}},{id:`openai/gpt-5.1-codex-mini`,name:`OpenAI: GPT-5.1-Codex-Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.1`,description:`GPT-5.1-Codex-Mini is a smaller and faster version of GPT-5.1-Codex`,context_window:4e5,max_output_tokens:1e5,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.25,output:2,cached_input:.025}},{id:`openai/gpt-5.1-codex`,name:`OpenAI: GPT-5.1-Codex`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.1`,description:`GPT-5.1-Codex is a specialized version of GPT-5.1 optimized for software engineering and coding workflows.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.25,output:10,cached_input:.125}},{id:`openai/gpt-5.1`,name:`OpenAI: GPT-5.1`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.1`,description:`GPT-5.1 is the latest frontier-grade model in the GPT-5 series, offering stronger general-purpose reasoning, improved instruction adherence, and a more natural conversational style compared to GPT-5.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.25,output:10,cached_input:.125}},{id:`openai/gpt-5.2-chat`,name:`OpenAI: GPT-5.2 Chat`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.2`,description:`GPT-5.2 Chat (AKA Instant) is the fast, lightweight member of the 5.2 family, optimized for low-latency chat while retaining strong general intelligence.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:1.75,output:14,cached_input:.175}},{id:`openai/gpt-5.2-codex`,name:`OpenAI: GPT-5.2-Codex`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.2`,description:`GPT-5.2-Codex is an upgraded version of GPT-5.1-Codex optimized for software engineering and coding workflows.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.75,output:14,cached_input:.175}},{id:`openai/gpt-5.2-pro`,name:`OpenAI: GPT-5.2 Pro`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.2`,description:`GPT-5.2 Pro is OpenAI’s most advanced model, offering major improvements in agentic coding and long context performance over GPT-5 Pro.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:21,output:168}},{id:`openai/gpt-5.2`,name:`OpenAI: GPT-5.2`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.2`,description:`GPT-5.2 is the latest frontier-grade model in the GPT-5 series, offering stronger agentic and long context perfomance compared to GPT-5.1.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.75,output:14,cached_input:.175}},{id:`openai/gpt-5.3-chat`,name:`OpenAI: GPT-5.3 Chat`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.3`,description:`GPT-5.3 Chat is an update to ChatGPT's most-used model that makes everyday conversations smoother, more useful, and more directly helpful.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:1.75,output:14,cached_input:.175}},{id:`openai/gpt-5.3-codex`,name:`OpenAI: GPT-5.3-Codex`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.3`,description:`GPT-5.3-Codex is OpenAI’s most advanced agentic coding model, combining the frontier software engineering performance of GPT-5.2-Codex with the broader reasoning and professional knowledge capabilities of GPT-5.2.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.75,output:14,cached_input:.175}},{id:`openai/gpt-5.4-mini`,name:`OpenAI: GPT-5.4 Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.4`,description:`GPT-5.4 mini brings the core capabilities of GPT-5.4 to a faster, more efficient model optimized for high-throughput workloads.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.75,output:4.5,cached_input:.075}},{id:`openai/gpt-5.4-nano`,name:`OpenAI: GPT-5.4 Nano`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.4`,description:`GPT-5.4 nano is the most lightweight and cost-efficient variant of the GPT-5.4 family, optimized for speed-critical and high-volume tasks.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.2,output:1.25,cached_input:.02}},{id:`openai/gpt-5.4-pro`,name:`OpenAI: GPT-5.4 Pro`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.4`,description:`GPT-5.4 Pro is OpenAI's most advanced model, building on GPT-5.4's unified architecture with enhanced reasoning capabilities for complex, high-stakes tasks.`,context_window:105e4,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:30,output:180}},{id:`openai/gpt-5.4`,name:`OpenAI: GPT-5.4`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5.4`,description:`GPT-5.4 is OpenAI’s latest frontier model, unifying the Codex and GPT lines into a single system.`,context_window:105e4,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:2.5,output:15,cached_input:.25}},{id:`openai/gpt-5`,name:`OpenAI: GPT-5`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-5`,description:`GPT-5 is OpenAI’s most advanced model, offering major improvements in reasoning, code quality, and user experience.`,context_window:4e5,max_output_tokens:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.25,output:10,cached_input:.125}},{id:`openai/gpt-audio-mini`,name:`OpenAI: GPT Audio Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-audio`,description:`A cost-efficient version of GPT Audio.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`audio`],output:[`text`,`audio`]},capabilities:{structured_output:!0},pricing:{input:.6,output:2.4}},{id:`openai/gpt-audio`,name:`OpenAI: GPT Audio`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-audio`,description:`The gpt-audio model is OpenAI's first generally available audio model.`,context_window:128e3,max_output_tokens:16384,modalities:{input:[`text`,`audio`],output:[`text`,`audio`]},capabilities:{structured_output:!0},pricing:{input:2.5,output:10}},{id:`openai/gpt-oss-120b:free`,name:`OpenAI: gpt-oss-120b (free)`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-oss`,description:`gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0}},{id:`openai/gpt-oss-120b`,name:`OpenAI: gpt-oss-120b`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-oss`,description:`gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.039,output:.19}},{id:`openai/gpt-oss-20b:free`,name:`OpenAI: gpt-oss-20b (free)`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-oss`,description:`gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0}},{id:`openai/gpt-oss-20b`,name:`OpenAI: gpt-oss-20b`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-oss`,description:`gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.03,output:.11,cached_input:.015}},{id:`openai/gpt-oss-safeguard-20b`,name:`OpenAI: gpt-oss-safeguard-20b`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`gpt-oss`,description:`gpt-oss-safeguard-20b is a safety reasoning model from OpenAI built upon gpt-oss-20b.`,context_window:131072,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.075,output:.3,cached_input:.037}},{id:`openai/o1-pro`,name:`OpenAI: o1-pro`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`o1`,description:`The o1 series of models are trained with reinforcement learning to think before they answer and perform complex reasoning.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:150,output:600}},{id:`openai/o1`,name:`OpenAI: o1`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`o1`,description:`The latest and strongest model family from OpenAI, o1 is designed to spend more time thinking before responding.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:15,output:60,cached_input:7.5}},{id:`openai/o3-deep-research`,name:`OpenAI: o3 Deep Research`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`o3`,description:`o3-deep-research is OpenAI's advanced model for deep research, designed to tackle complex, multi-step research tasks.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:10,output:40,cached_input:2.5}},{id:`openai/o3-mini-high`,name:`OpenAI: o3 Mini High`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`o3`,description:`OpenAI o3-mini-high is the same model as [o3-mini](/openai/o3-mini) with reasoning_effort set to high.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.1,output:4.4,cached_input:.55}},{id:`openai/o3-mini`,name:`OpenAI: o3 Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`o3`,description:`OpenAI o3-mini is a cost-efficient language model optimized for STEM reasoning tasks, particularly excelling in science, mathematics, and coding.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.1,output:4.4,cached_input:.55}},{id:`openai/o3-pro`,name:`OpenAI: o3 Pro`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`o3`,description:`The o-series of models are trained with reinforcement learning to think before they answer and perform complex reasoning.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:20,output:80}},{id:`openai/o3`,name:`OpenAI: o3`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`o3`,description:`o3 is a well-rounded and powerful model across domains.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:2,output:8,cached_input:.5}},{id:`openai/o4-mini-deep-research`,name:`OpenAI: o4 Mini Deep Research`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`o4`,description:`o4-mini-deep-research is OpenAI's faster, more affordable deep research model—ideal for tackling complex, multi-step research tasks.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:2,output:8,cached_input:.5}},{id:`openai/o4-mini-high`,name:`OpenAI: o4 Mini High`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`o4`,description:`OpenAI o4-mini-high is the same model as [o4-mini](/openai/o4-mini) with reasoning_effort set to high.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.1,output:4.4,cached_input:.275}},{id:`openai/o4-mini`,name:`OpenAI: o4 Mini`,created_by:`openai`,source:`official`,last_updated:`2026-03-21`,family:`o4`,description:`OpenAI o4-mini is a compact reasoning model in the o-series, optimized for fast, cost-efficient performance while retaining strong multimodal and agentic capabilities.`,context_window:2e5,max_output_tokens:1e5,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1.1,output:4.4,cached_input:.275}},{id:`openrouter/auto`,name:`Auto Router`,created_by:`openrouter`,source:`official`,last_updated:`2026-03-21`,description:`Your prompt will be processed by a meta-model and routed to one of dozens of models (see below), optimizing for the best possible output.`,context_window:2e6,modalities:{input:[`text`,`image`,`audio`,`video`],output:[`text`,`image`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0}},{id:`openrouter/bodybuilder`,name:`Body Builder (beta)`,created_by:`openrouter`,source:`official`,last_updated:`2026-03-21`,description:`Transform your natural language requests into structured OpenRouter API request objects.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]}},{id:`openrouter/free`,name:`Free Models Router`,created_by:`openrouter`,source:`official`,last_updated:`2026-03-21`,description:`The simplest way to get free inference.`,context_window:2e5,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0}},{id:`perplexity/sonar-deep-research`,name:`Perplexity: Sonar Deep Research`,created_by:`perplexity`,source:`official`,last_updated:`2026-03-21`,description:`Sonar Deep Research is a research-focused model designed for multi-step retrieval, synthesis, and reasoning across complex topics.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:2,output:8}},{id:`perplexity/sonar-pro-search`,name:`Perplexity: Sonar Pro Search`,created_by:`perplexity`,source:`official`,last_updated:`2026-03-21`,description:`Exclusively available on the OpenRouter API, Sonar Pro's new Pro Search mode is Perplexity's most advanced agentic search system.`,context_window:2e5,max_output_tokens:8e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:3,output:15}},{id:`perplexity/sonar-pro`,name:`Perplexity: Sonar Pro`,created_by:`perplexity`,source:`official`,last_updated:`2026-03-21`,description:`Note: Sonar Pro pricing includes Perplexity search pricing.`,context_window:2e5,max_output_tokens:8e3,modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:3,output:15}},{id:`perplexity/sonar-reasoning-pro`,name:`Perplexity: Sonar Reasoning Pro`,created_by:`perplexity`,source:`official`,last_updated:`2026-03-21`,description:`Note: Sonar Pro pricing includes Perplexity search pricing.`,context_window:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:2,output:8}},{id:`perplexity/sonar`,name:`Perplexity: Sonar`,created_by:`perplexity`,source:`official`,last_updated:`2026-03-21`,description:`Sonar is lightweight, affordable, fast, and simple to use — now featuring citations and the ability to customize sources.`,context_window:127072,modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:1,output:1}},{id:`prime-intellect/intellect-3`,name:`Prime Intellect: INTELLECT-3`,created_by:`prime-intellect`,source:`official`,last_updated:`2026-03-21`,description:`INTELLECT-3 is a 106B-parameter Mixture-of-Experts model (12B active) post-trained from GLM-4.5-Air-Base using supervised fine-tuning (SFT) followed by large-scale reinforcement learning (RL).`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.2,output:1.1}},{id:`qwen/qwen-2.5-72b-instruct`,name:`Qwen2.5 72B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen`,description:`Qwen2.5 72B is the latest series of Qwen large language models.`,context_window:32768,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.12,output:.39}},{id:`qwen/qwen-2.5-7b-instruct`,name:`Qwen: Qwen2.5 7B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen`,description:`Qwen2.5 7B is the latest series of Qwen large language models.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.04,output:.1}},{id:`qwen/qwen-2.5-coder-32b-instruct`,name:`Qwen2.5 Coder 32B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen`,description:`Qwen2.5-Coder is the latest series of Code-Specific Qwen large language models (formerly known as CodeQwen).`,context_window:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.66,output:1}},{id:`qwen/qwen-2.5-vl-7b-instruct`,name:`Qwen: Qwen2.5-VL 7B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen`,description:`Qwen2.5 VL 7B is a multimodal LLM from the Qwen Team with the following key enhancements:
+- SoTA understanding of images of various resolution & ratio: Qwen2.5-VL achieves state-of-the-art performanc`},{id:`qwen/qwen-max`,name:`Qwen: Qwen-Max `,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen`,description:`Qwen-Max, based on Qwen2.5, provides the best inference performance among [Qwen models](/qwen), especially for complex multi-step tasks.`,context_window:32768,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:1.04,output:4.16,cached_input:.208},model_type:`chat`,release_date:`2025-02-01`,tagline:`Qwen-Max, based on Qwen2.5, provides the best inference performance among [Qwen models](/qwen), especially for complex multi-step tasks.`,tools:[`function_calling`]},{id:`qwen/qwen-plus-2025-07-28:thinking`,name:`Qwen: Qwen Plus 0728 (thinking)`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen`,description:`Qwen Plus 0728, based on the Qwen3 foundation model, is a 1 million context hybrid reasoning model with a balanced performance, speed, and cost combination.`,context_window:1e6,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.26,output:.78},model_type:`chat`,release_date:`2025-09-08`,tagline:`Qwen Plus 0728, based on the Qwen3 foundation model, is a 1 million context hybrid reasoning model with a balanced performance, speed, and cost combination.`,tools:[`function_calling`]},{id:`qwen/qwen-plus-2025-07-28`,name:`Qwen: Qwen Plus 0728`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen`,description:`Qwen Plus 0728, based on the Qwen3 foundation model, is a 1 million context hybrid reasoning model with a balanced performance, speed, and cost combination.`,context_window:1e6,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.26,output:.78,cached_input:.052},model_type:`chat`,release_date:`2025-09-08`,alias:`qwen/qwen-plus`,tagline:`Qwen Plus 0728, based on the Qwen3 foundation model, is a 1 million context hybrid reasoning model with a balanced performance, speed, and cost combination.`,tools:[`function_calling`]},{id:`qwen/qwen-plus`,name:`Qwen: Qwen-Plus`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen`,description:`Qwen-Plus, based on the Qwen2.5 foundation model, is a 131K context model with a balanced performance, speed, and cost combination.`,context_window:1e6,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,reasoning:!0},pricing:{input:.26,output:.78,cached_input:.052},model_type:`chat`,release_date:`2025-02-01`,tagline:`Qwen Plus 0728, based on the Qwen3 foundation model, is a 1 million context hybrid reasoning model with a balanced performance, speed, and cost combination.`,tools:[`function_calling`],snapshots:[`qwen/qwen-plus-2025-07-28`],license:`proprietary`},{id:`qwen/qwen-turbo`,name:`Qwen: Qwen-Turbo`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen`,description:`Qwen-Turbo, based on Qwen2.5, is a 1M context model that provides fast speed and low cost, suitable for simple tasks.`,context_window:131072,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,reasoning:!0},pricing:{input:.033,output:.13,cached_input:.007},model_type:`chat`,release_date:`2025-02-01`,tagline:`Qwen-Turbo, based on Qwen2.5, is a 1M context model that provides fast speed and low cost, suitable for simple tasks.`,tools:[`function_calling`],license:`proprietary`},{id:`qwen/qwen-vl-max`,name:`Qwen: Qwen VL Max`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen`,description:`Qwen VL Max is a visual understanding model with 7500 tokens context length.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.52,output:2.08},model_type:`chat`,release_date:`2025-02-01`,tagline:`Qwen VL Max is a visual understanding model with 7500 tokens context length.`,tools:[`function_calling`]},{id:`qwen/qwen-vl-plus`,name:`Qwen: Qwen VL Plus`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen`,description:`Qwen's Enhanced Large Visual Language Model.`,context_window:131072,max_output_tokens:8192,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.137,output:.41,cached_input:.027},model_type:`chat`,release_date:`2025-02-05`,tagline:`Qwen's Enhanced Large Visual Language Model.`},{id:`qwen/qwen2.5-coder-7b-instruct`,name:`Qwen: Qwen2.5 Coder 7B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen2.5`,description:`Qwen2.5-Coder-7B-Instruct is a 7B parameter instruction-tuned language model optimized for code-related tasks such as code generation, reasoning, and bug fixing.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.03,output:.09},model_type:`code`,parameters:7,release_date:`2025-04-15`,tagline:`Qwen2.5-Coder-7B-Instruct is a 7B parameter instruction-tuned language model optimized for code-related tasks such as code generation, reasoning, and bug fixing.`},{id:`qwen/qwen2.5-vl-32b-instruct`,name:`Qwen: Qwen2.5 VL 32B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen2.5`,description:`Qwen2.5-VL-32B is a multimodal vision-language model fine-tuned through reinforcement learning for enhanced mathematical reasoning, structured outputs, and visual problem-solving capabilities.`,context_window:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.2,output:.6},model_type:`chat`,parameters:32,release_date:`2025-03-24`,tagline:`Qwen2.5-VL-32B is a multimodal vision-language model fine-tuned through reinforcement learning for enhanced mathematical reasoning, structured outputs, and visual problem-solving capabilities.`},{id:`qwen/qwen2.5-vl-72b-instruct`,name:`Qwen: Qwen2.5 VL 72B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen2.5`,description:`Qwen2.5-VL is proficient in recognizing common objects such as flowers, birds, fish, and insects.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.8,output:.8},model_type:`chat`,parameters:72,release_date:`2025-02-01`,tagline:`Qwen2.5-VL is proficient in recognizing common objects such as flowers, birds, fish, and insects.`},{id:`qwen/qwen3-14b`,name:`Qwen: Qwen3 14B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-14B is a dense 14.8B parameter causal language model from the Qwen3 series, designed for both complex reasoning and efficient dialogue.`,context_window:40960,max_output_tokens:40960,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.06,output:.24},model_type:`chat`,parameters:14,release_date:`2025-04-28`,tagline:`Qwen3-14B is a dense 14.8B parameter causal language model from the Qwen3 series, designed for both complex reasoning and efficient dialogue.`,tools:[`function_calling`]},{id:`qwen/qwen3-235b-a22b-2507`,name:`Qwen: Qwen3 235B A22B Instruct 2507`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-235B-A22B-Instruct-2507 is a multilingual, instruction-tuned mixture-of-experts language model based on the Qwen3-235B architecture, with 22B active parameters per forward pass.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.071,output:.1},model_type:`chat`,parameters:235,active_parameters:22,release_date:`2025-07-21`,max_output_tokens:8192,alias:`qwen/qwen3-235b-a22b`,tagline:`Qwen3-235B-A22B-Instruct-2507 is a multilingual, instruction-tuned mixture-of-experts language model based on the Qwen3-235B architecture, with 22B active parameters per forward pass.`,tools:[`function_calling`]},{id:`qwen/qwen3-235b-a22b-thinking-2507`,name:`Qwen: Qwen3 235B A22B Thinking 2507`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.15,output:1.495},model_type:`chat`,parameters:235,active_parameters:22,release_date:`2025-07-25`,alias:`qwen/qwen3-235b-a22b-thinking`,tagline:`Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-235b-a22b-thinking`,name:`Qwen: Qwen3 235B A22B Thinking 2507`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks.`,release_date:`2025-07-25`,context_window:131072,parameters:235,active_parameters:22,model_type:`chat`,tagline:`Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks.`,capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},modalities:{input:[`text`],output:[`text`]},pricing:{input:.15,output:1.495},tools:[`function_calling`],snapshots:[`qwen/qwen3-235b-a22b-thinking-2507`],license:`apache-2.0`},{id:`qwen/qwen3-235b-a22b`,name:`Qwen: Qwen3 235B A22B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-235B-A22B is a 235B parameter mixture-of-experts (MoE) model developed by Qwen, activating 22B parameters per forward pass.`,context_window:131072,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.455,output:1.82},model_type:`chat`,parameters:235,active_parameters:22,release_date:`2025-04-28`,tagline:`Qwen3-235B-A22B-Instruct-2507 is a multilingual, instruction-tuned mixture-of-experts language model based on the Qwen3-235B architecture, with 22B active parameters per forward pass.`,tools:[`function_calling`],snapshots:[`qwen/qwen3-235b-a22b-2507`],license:`apache-2.0`},{id:`qwen/qwen3-30b-a3b-instruct-2507`,name:`Qwen: Qwen3 30B A3B Instruct 2507`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-30B-A3B-Instruct-2507 is a 30.5B-parameter mixture-of-experts language model from Qwen, with 3.3B active parameters per inference.`,context_window:262144,max_output_tokens:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:.09,output:.3},model_type:`chat`,parameters:30,active_parameters:3,release_date:`2025-07-29`,alias:`qwen/qwen3-30b-a3b-instruct`,tagline:`Qwen3-30B-A3B-Instruct-2507 is a 30.5B-parameter mixture-of-experts language model from Qwen, with 3.3B active parameters per inference.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-30b-a3b-instruct`,name:`Qwen: Qwen3 30B A3B Instruct 2507`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-30B-A3B-Instruct-2507 is a 30.5B-parameter mixture-of-experts language model from Qwen, with 3.3B active parameters per inference.`,release_date:`2025-07-29`,context_window:262144,max_output_tokens:262144,parameters:30,active_parameters:3,model_type:`chat`,tagline:`Qwen3-30B-A3B-Instruct-2507 is a 30.5B-parameter mixture-of-experts language model from Qwen, with 3.3B active parameters per inference.`,capabilities:{tool_call:!0,structured_output:!0,streaming:!0},modalities:{input:[`text`],output:[`text`]},pricing:{input:.09,output:.3},tools:[`function_calling`],snapshots:[`qwen/qwen3-30b-a3b-instruct-2507`],license:`apache-2.0`},{id:`qwen/qwen3-30b-a3b-thinking-2507`,name:`Qwen: Qwen3 30B A3B Thinking 2507`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-30B-A3B-Thinking-2507 is a 30B parameter Mixture-of-Experts reasoning model optimized for complex tasks requiring extended multi-step thinking.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.08,output:.4,cached_input:.08},model_type:`chat`,parameters:30,active_parameters:3,release_date:`2025-08-28`,alias:`qwen/qwen3-30b-a3b-thinking`,tagline:`Qwen3-30B-A3B-Thinking-2507 is a 30B parameter Mixture-of-Experts reasoning model optimized for complex tasks requiring extended multi-step thinking.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-30b-a3b-thinking`,name:`Qwen: Qwen3 30B A3B Thinking 2507`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-30B-A3B-Thinking-2507 is a 30B parameter Mixture-of-Experts reasoning model optimized for complex tasks requiring extended multi-step thinking.`,release_date:`2025-08-28`,context_window:131072,max_output_tokens:131072,parameters:30,active_parameters:3,model_type:`chat`,tagline:`Qwen3-30B-A3B-Thinking-2507 is a 30B parameter Mixture-of-Experts reasoning model optimized for complex tasks requiring extended multi-step thinking.`,capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},modalities:{input:[`text`],output:[`text`]},pricing:{input:.08,output:.4,cached_input:.08},tools:[`function_calling`],snapshots:[`qwen/qwen3-30b-a3b-thinking-2507`],license:`apache-2.0`},{id:`qwen/qwen3-30b-a3b`,name:`Qwen: Qwen3 30B A3B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3, the latest generation in the Qwen large language model series, features both dense and mixture-of-experts (MoE) architectures to excel in reasoning, multilingual support, and advanced agent tasks.`,context_window:40960,max_output_tokens:40960,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.08,output:.28},model_type:`chat`,parameters:30,active_parameters:3,release_date:`2025-04-28`,tagline:`Qwen3, the latest generation in the Qwen large language model series, features both dense and mixture-of-experts (MoE) architectures to excel in reasoning, multilingual support, and advanced agent tas`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-32b`,name:`Qwen: Qwen3 32B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-32B is a dense 32.8B parameter causal language model from the Qwen3 series, optimized for both complex reasoning and efficient dialogue.`,context_window:40960,max_output_tokens:40960,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.08,output:.24,cached_input:.04},model_type:`chat`,parameters:32,release_date:`2025-04-28`,tagline:`Qwen3-32B is a dense 32.8B parameter causal language model from the Qwen3 series, optimized for both complex reasoning and efficient dialogue.`,tools:[`function_calling`]},{id:`qwen/qwen3-4b:free`,name:`Qwen: Qwen3 4B (free)`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-4B is a 4 billion parameter dense language model from the Qwen3 series, designed to support both general-purpose and reasoning-intensive tasks.`,status:`deprecated`,deprecation_date:`2026-03-29`,context_window:40960,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},model_type:`chat`,release_date:`2025-04-30`,tagline:`Qwen3-4B is a 4 billion parameter dense language model from the Qwen3 series, designed to support both general-purpose and reasoning-intensive tasks.`,tools:[`function_calling`]},{id:`qwen/qwen3-8b`,name:`Qwen: Qwen3 8B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-8B is a dense 8.2B parameter causal language model from the Qwen3 series, designed for both reasoning-heavy tasks and efficient dialogue.`,context_window:40960,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.05,output:.4,cached_input:.05},model_type:`chat`,parameters:8,release_date:`2025-04-28`,tagline:`Qwen3-8B is a dense 8.2B parameter causal language model from the Qwen3 series, designed for both reasoning-heavy tasks and efficient dialogue.`,tools:[`function_calling`]},{id:`qwen/qwen3-coder-30b-a3b-instruct`,name:`Qwen: Qwen3 Coder 30B A3B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-Coder-30B-A3B-Instruct is a 30.5B parameter Mixture-of-Experts (MoE) model with 128 experts (8 active per forward pass), designed for advanced code generation, repository-scale understanding, and agentic tool use.`,context_window:16e4,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:.07,output:.27},model_type:`code`,parameters:30,active_parameters:3,release_date:`2025-07-31`,tagline:`Qwen3-Coder-30B-A3B-Instruct is a 30.5B parameter Mixture-of-Experts (MoE) model with 128 experts (8 active per forward pass), designed for advanced code generation, repository-scale understanding, an`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-coder-flash`,name:`Qwen: Qwen3 Coder Flash`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3 Coder Flash is Alibaba's fast and cost efficient version of their proprietary Qwen3 Coder Plus.`,context_window:1e6,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:.195,output:.975,cached_input:.039},model_type:`code`,release_date:`2025-09-17`,tagline:`Qwen3 Coder Flash is Alibaba's fast and cost efficient version of their proprietary Qwen3 Coder Plus.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-coder:free`,name:`Qwen: Qwen3 Coder 480B A35B (free)`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-Coder-480B-A35B-Instruct is a Mixture-of-Experts (MoE) code generation model developed by the Qwen team.`,context_window:262e3,max_output_tokens:262e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},model_type:`code`,parameters:480,active_parameters:35,release_date:`2025-07-23`,tagline:`Qwen3-Coder-480B-A35B-Instruct is a Mixture-of-Experts (MoE) code generation model developed by the Qwen team.`,tools:[`function_calling`]},{id:`qwen/qwen3-coder-next`,name:`Qwen: Qwen3 Coder Next`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-Coder-Next is an open-weight causal language model optimized for coding agents and local development workflows.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:.12,output:.75,cached_input:.06},model_type:`code`,release_date:`2026-02-04`,tagline:`Qwen3-Coder-Next is an open-weight causal language model optimized for coding agents and local development workflows.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-coder-plus`,name:`Qwen: Qwen3 Coder Plus`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3 Coder Plus is Alibaba's proprietary version of the Open Source Qwen3 Coder 480B A35B.`,context_window:1e6,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:.65,output:3.25,cached_input:.13},model_type:`code`,release_date:`2025-09-23`,tagline:`Qwen3 Coder Plus is Alibaba's proprietary version of the Open Source Qwen3 Coder 480B A35B.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-coder`,name:`Qwen: Qwen3 Coder 480B A35B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-Coder-480B-A35B-Instruct is a Mixture-of-Experts (MoE) code generation model developed by the Qwen team.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.22,output:1,cached_input:.022},model_type:`code`,parameters:480,active_parameters:35,release_date:`2025-07-23`,tagline:`Qwen3-Coder-480B-A35B-Instruct is a Mixture-of-Experts (MoE) code generation model developed by the Qwen team.`,tools:[`function_calling`]},{id:`qwen/qwen3-max-thinking`,name:`Qwen: Qwen3 Max Thinking`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-Max-Thinking is the flagship reasoning model in the Qwen3 series, designed for high-stakes cognitive tasks that require deep, multi-step reasoning.`,context_window:262144,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.78,output:3.9},model_type:`chat`,release_date:`2026-02-09`,tagline:`Qwen3-Max-Thinking is the flagship reasoning model in the Qwen3 series, designed for high-stakes cognitive tasks that require deep, multi-step reasoning.`,tools:[`function_calling`]},{id:`qwen/qwen3-max`,name:`Qwen: Qwen3 Max`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-Max is an updated release built on the Qwen3 series, offering major improvements in reasoning, instruction following, multilingual support, and long-tail knowledge coverage compared to the January 2025 version.`,context_window:262144,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,reasoning:!0},pricing:{input:.78,output:3.9,cached_input:.156},model_type:`chat`,release_date:`2025-09-23`,tagline:`Qwen3-Max is an updated release built on the Qwen3 series, offering major improvements in reasoning, instruction following, multilingual support, and long-tail knowledge coverage compared to the Janua`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-next-80b-a3b-instruct:free`,name:`Qwen: Qwen3 Next 80B A3B Instruct (free)`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-Next-80B-A3B-Instruct is an instruction-tuned chat model in the Qwen3-Next series optimized for fast, stable responses without “thinking” traces.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},model_type:`chat`,parameters:80,active_parameters:3,release_date:`2025-09-11`,tagline:`Qwen3-Next-80B-A3B-Instruct is an instruction-tuned chat model in the Qwen3-Next series optimized for fast, stable responses without “thinking” traces.`,tools:[`function_calling`]},{id:`qwen/qwen3-next-80b-a3b-instruct`,name:`Qwen: Qwen3 Next 80B A3B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-Next-80B-A3B-Instruct is an instruction-tuned chat model in the Qwen3-Next series optimized for fast, stable responses without “thinking” traces.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:.09,output:1.1},model_type:`chat`,parameters:80,active_parameters:3,release_date:`2025-09-11`,tagline:`Qwen3-Next-80B-A3B-Instruct is an instruction-tuned chat model in the Qwen3-Next series optimized for fast, stable responses without “thinking” traces.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-next-80b-a3b-thinking`,name:`Qwen: Qwen3 Next 80B A3B Thinking`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-Next-80B-A3B-Thinking is a reasoning-first chat model in the Qwen3-Next line that outputs structured “thinking” traces by default.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.098,output:.78},model_type:`chat`,parameters:80,active_parameters:3,release_date:`2025-09-11`,tagline:`Qwen3-Next-80B-A3B-Thinking is a reasoning-first chat model in the Qwen3-Next line that outputs structured “thinking” traces by default.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-vl-235b-a22b-instruct`,name:`Qwen: Qwen3 VL 235B A22B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-VL-235B-A22B Instruct is an open-weight multimodal model that unifies strong text generation with visual understanding across images and video.`,context_window:262144,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0,reasoning:!0},pricing:{input:.2,output:.88,cached_input:.11},model_type:`chat`,parameters:235,active_parameters:22,release_date:`2025-09-23`,tagline:`Qwen3-VL-235B-A22B Instruct is an open-weight multimodal model that unifies strong text generation with visual understanding across images and video.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-vl-235b-a22b-thinking`,name:`Qwen: Qwen3 VL 235B A22B Thinking`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-VL-235B-A22B Thinking is a multimodal model that unifies strong text generation with visual understanding across images and video.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:.26,output:2.6},model_type:`chat`,parameters:235,active_parameters:22,release_date:`2025-09-23`,tagline:`Qwen3-VL-235B-A22B Thinking is a multimodal model that unifies strong text generation with visual understanding across images and video.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-vl-30b-a3b-instruct`,name:`Qwen: Qwen3 VL 30B A3B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-VL-30B-A3B-Instruct is a multimodal model that unifies strong text generation with visual understanding for images and videos.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0,reasoning:!0},pricing:{input:.13,output:.52},model_type:`chat`,parameters:30,active_parameters:3,release_date:`2025-10-06`,tagline:`Qwen3-VL-30B-A3B-Instruct is a multimodal model that unifies strong text generation with visual understanding for images and videos.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-vl-30b-a3b-thinking`,name:`Qwen: Qwen3 VL 30B A3B Thinking`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-VL-30B-A3B-Thinking is a multimodal model that unifies strong text generation with visual understanding for images and videos.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:.13,output:1.56},model_type:`chat`,parameters:30,active_parameters:3,release_date:`2025-10-06`,tagline:`Qwen3-VL-30B-A3B-Thinking is a multimodal model that unifies strong text generation with visual understanding for images and videos.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-vl-32b-instruct`,name:`Qwen: Qwen3 VL 32B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-VL-32B-Instruct is a large-scale multimodal vision-language model designed for high-precision understanding and reasoning across text, images, and video.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0,reasoning:!0},pricing:{input:.104,output:.416},model_type:`chat`,parameters:32,release_date:`2025-10-23`,tagline:`Qwen3-VL-32B-Instruct is a large-scale multimodal vision-language model designed for high-precision understanding and reasoning across text, images, and video.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-vl-8b-instruct`,name:`Qwen: Qwen3 VL 8B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-VL-8B-Instruct is a multimodal vision-language model from the Qwen3-VL series, built for high-fidelity understanding and reasoning across text, images, and video.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0,vision:!0,reasoning:!0},pricing:{input:.08,output:.5},model_type:`chat`,parameters:8,release_date:`2025-10-14`,tagline:`Qwen3-VL-8B-Instruct is a multimodal vision-language model from the Qwen3-VL series, built for high-fidelity understanding and reasoning across text, images, and video.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3-vl-8b-thinking`,name:`Qwen: Qwen3 VL 8B Thinking`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3`,description:`Qwen3-VL-8B-Thinking is the reasoning-optimized variant of the Qwen3-VL-8B multimodal model, designed for advanced visual and textual reasoning across complex scenes, documents, and temporal sequences.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0,vision:!0},pricing:{input:.117,output:1.365},model_type:`chat`,parameters:8,release_date:`2025-10-14`,tagline:`Qwen3-VL-8B-Thinking is the reasoning-optimized variant of the Qwen3-VL-8B multimodal model, designed for advanced visual and textual reasoning across complex scenes, documents, and temporal sequences`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3.5-122b-a10b`,name:`Qwen: Qwen3.5-122B-A10B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3.5`,description:`The Qwen3.5 122B-A10B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.26,output:2.08},model_type:`chat`,parameters:122,active_parameters:10,release_date:`2026-02-25`,tagline:`The Qwen3.5 122B-A10B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference eff`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3.5-27b`,name:`Qwen: Qwen3.5-27B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3.5`,description:`The Qwen3.5 27B native vision-language Dense model incorporates a linear attention mechanism, delivering fast response times while balancing inference speed and performance.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.195,output:1.56},model_type:`chat`,parameters:27,release_date:`2026-02-25`,tagline:`The Qwen3.5 27B native vision-language Dense model incorporates a linear attention mechanism, delivering fast response times while balancing inference speed and performance.`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3.5-35b-a3b`,name:`Qwen: Qwen3.5-35B-A3B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3.5`,description:`The Qwen3.5 Series 35B-A3B is a native vision-language model designed with a hybrid architecture that integrates linear attention mechanisms and a sparse mixture-of-experts model, achieving higher inference efficiency.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.163,output:1.3},model_type:`chat`,parameters:35,active_parameters:3,release_date:`2026-02-25`,tagline:`The Qwen3.5 Series 35B-A3B is a native vision-language model designed with a hybrid architecture that integrates linear attention mechanisms and a sparse mixture-of-experts model, achieving higher inf`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3.5-397b-a17b`,name:`Qwen: Qwen3.5 397B A17B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3.5`,description:`The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.39,output:2.34},model_type:`chat`,parameters:397,active_parameters:17,release_date:`2026-02-16`,tagline:`The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher infere`,tools:[`function_calling`],license:`apache-2.0`},{id:`qwen/qwen3.5-9b`,name:`Qwen: Qwen3.5-9B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3.5`,description:`Qwen3.5-9B is a multimodal foundation model from the Qwen3.5 family, designed to deliver strong reasoning, coding, and visual understanding in an efficient 9B-parameter architecture.`,context_window:256e3,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.05,output:.15},model_type:`chat`,parameters:9,release_date:`2026-03-10`,tagline:`Qwen3.5-9B is a multimodal foundation model from the Qwen3.5 family, designed to deliver strong reasoning, coding, and visual understanding in an efficient 9B-parameter architecture.`,tools:[`function_calling`]},{id:`qwen/qwen3.5-flash-02-23`,name:`Qwen: Qwen3.5-Flash`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3.5`,description:`The Qwen3.5 native vision-language Flash models are built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency.`,context_window:1e6,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.065,output:.26},model_type:`chat`,release_date:`2026-02-25`,tagline:`The Qwen3.5 native vision-language Flash models are built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference effic`,tools:[`function_calling`]},{id:`qwen/qwen3.5-plus-02-15`,name:`Qwen: Qwen3.5 Plus 2026-02-15`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwen3.5`,description:`The Qwen3.5 native vision-language series Plus models are built on a hybrid architecture that integrates linear attention mechanisms with sparse mixture-of-experts models, achieving higher inference efficiency.`,context_window:1e6,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.26,output:1.56},model_type:`chat`,release_date:`2026-02-16`,tagline:`The Qwen3.5 native vision-language series Plus models are built on a hybrid architecture that integrates linear attention mechanisms with sparse mixture-of-experts models, achieving higher inference e`,tools:[`function_calling`]},{id:`qwen/qwq-32b`,name:`Qwen: QwQ 32B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-23`,family:`qwq`,description:`QwQ is the reasoning model of the Qwen series.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0,streaming:!0},pricing:{input:.15,output:.58},model_type:`chat`,parameters:32,release_date:`2025-03-05`,tagline:`QwQ is the reasoning model of the Qwen series.`,tools:[`function_calling`],license:`apache-2.0`},{id:`relace/relace-apply-3`,name:`Relace: Relace Apply 3`,created_by:`relace`,source:`official`,last_updated:`2026-03-23`,description:`Relace Apply 3 is a specialized code-patching LLM that merges AI-suggested edits straight into your source files.`,context_window:256e3,max_output_tokens:128e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:.85,output:1.25},release_date:`2025-09-26`,tagline:`Relace Apply 3 is a specialized code-patching LLM that merges AI-suggested edits straight into your source files.`},{id:`relace/relace-search`,name:`Relace: Relace Search`,created_by:`relace`,source:`official`,last_updated:`2026-03-23`,description:"The relace-search model uses 4-12 `view_file` and `grep` tools in parallel to explore a codebase and return relevant files to the user request.",context_window:256e3,max_output_tokens:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:1,output:3},model_type:`chat`,release_date:`2025-12-08`,tagline:"The relace-search model uses 4-12 `view_file` and `grep` tools in parallel to explore a codebase and return relevant files to the user request.",tools:[`function_calling`]},{id:`sao10k/l3-euryale-70b`,name:`Sao10k: Llama 3 Euryale 70B v2.1`,created_by:`sao10k`,source:`official`,last_updated:`2026-03-23`,description:`Euryale 70B v2.1 is a model focused on creative roleplay from [Sao10k](https://ko-fi.com/sao10k).`,context_window:8192,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:1.48,output:1.48},parameters:70,release_date:`2024-06-18`,tagline:`Euryale 70B v2.1 is a model focused on creative roleplay from [Sao10k](https://ko-fi.com/sao10k).`,tools:[`function_calling`]},{id:`sao10k/l3-lunaris-8b`,name:`Sao10K: Llama 3 8B Lunaris`,created_by:`sao10k`,source:`official`,last_updated:`2026-03-23`,description:`Lunaris 8B is a versatile generalist and roleplaying model based on Llama 3.`,context_window:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.04,output:.05},parameters:8,release_date:`2024-08-13`,tagline:`Lunaris 8B is a versatile generalist and roleplaying model based on Llama 3.`},{id:`sao10k/l3.1-70b-hanami-x1`,name:`Sao10K: Llama 3.1 70B Hanami x1`,created_by:`sao10k`,source:`official`,last_updated:`2026-03-23`,description:`This is [Sao10K](/sao10k)'s experiment over [Euryale v2.2](/sao10k/l3.1-euryale-70b).`,context_window:16e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:3,output:3},parameters:70,release_date:`2025-01-08`,tagline:`This is [Sao10K](/sao10k)'s experiment over [Euryale v2.2](/sao10k/l3.1-euryale-70b).`},{id:`sao10k/l3.1-euryale-70b`,name:`Sao10K: Llama 3.1 Euryale 70B v2.2`,created_by:`sao10k`,source:`official`,last_updated:`2026-03-23`,description:`Euryale L3.1 70B v2.2 is a model focused on creative roleplay from [Sao10k](https://ko-fi.com/sao10k).`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.85,output:.85},parameters:70,release_date:`2024-08-28`,tagline:`Euryale L3.1 70B v2.2 is a model focused on creative roleplay from [Sao10k](https://ko-fi.com/sao10k).`,tools:[`function_calling`]},{id:`sao10k/l3.3-euryale-70b`,name:`Sao10K: Llama 3.3 Euryale 70B`,created_by:`sao10k`,source:`official`,last_updated:`2026-03-23`,description:`Euryale L3.3 70B is a model focused on creative roleplay from [Sao10k](https://ko-fi.com/sao10k).`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.65,output:.75},parameters:70,release_date:`2024-12-18`,tagline:`Euryale L3.3 70B is a model focused on creative roleplay from [Sao10k](https://ko-fi.com/sao10k).`},{id:`stepfun/step-3.5-flash:free`,name:`StepFun: Step 3.5 Flash (free)`,created_by:`stepfun`,source:`official`,last_updated:`2026-03-23`,description:`Step 3.5 Flash is StepFun's most capable open-source foundation model.`,context_window:256e3,max_output_tokens:256e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},release_date:`2026-01-29`,tagline:`Step 3.5 Flash is StepFun's most capable open-source foundation model.`,tools:[`function_calling`]},{id:`stepfun/step-3.5-flash`,name:`StepFun: Step 3.5 Flash`,created_by:`stepfun`,source:`official`,last_updated:`2026-03-23`,description:`Step 3.5 Flash is StepFun's most capable open-source foundation model.`,context_window:256e3,max_output_tokens:256e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:.1,output:.3,cached_input:.02},release_date:`2026-01-29`,tagline:`Step 3.5 Flash is StepFun's most capable open-source foundation model.`,tools:[`function_calling`]},{id:`switchpoint/router`,name:`Switchpoint Router`,created_by:`switchpoint`,source:`official`,last_updated:`2026-03-23`,description:`Switchpoint AI's router instantly analyzes your request and directs it to the optimal AI from an ever-evolving library.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:.85,output:3.4},release_date:`2025-07-11`,tagline:`Switchpoint AI's router instantly analyzes your request and directs it to the optimal AI from an ever-evolving library.`},{id:`tencent/hunyuan-a13b-instruct`,name:`Tencent: Hunyuan A13B Instruct`,created_by:`tencent`,source:`official`,last_updated:`2026-03-23`,description:`Hunyuan-A13B is a 13B active parameter Mixture-of-Experts (MoE) language model developed by Tencent, with a total parameter count of 80B and support for reasoning via Chain-of-Thought.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.14,output:.57},release_date:`2025-07-08`,tagline:`Hunyuan-A13B is a 13B active parameter Mixture-of-Experts (MoE) language model developed by Tencent, with a total parameter count of 80B and support for reasoning via Chain-of-Thought.`},{id:`thedrummer/cydonia-24b-v4.1`,name:`TheDrummer: Cydonia 24B V4.1`,created_by:`thedrummer`,source:`official`,last_updated:`2026-03-23`,description:`Uncensored and creative writing model based on Mistral Small 3.2 24B with good recall, prompt adherence, and intelligence.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},pricing:{input:.3,output:.5},parameters:24,release_date:`2025-09-27`,tagline:`Uncensored and creative writing model based on Mistral Small 3.2 24B with good recall, prompt adherence, and intelligence.`},{id:`thedrummer/rocinante-12b`,name:`TheDrummer: Rocinante 12B`,created_by:`thedrummer`,source:`official`,last_updated:`2026-03-23`,description:`Rocinante 12B is designed for engaging storytelling and rich prose.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.17,output:.43},parameters:12,release_date:`2024-09-30`,tagline:`Rocinante 12B is designed for engaging storytelling and rich prose.`,tools:[`function_calling`]},{id:`thedrummer/skyfall-36b-v2`,name:`TheDrummer: Skyfall 36B V2`,created_by:`thedrummer`,source:`official`,last_updated:`2026-03-23`,description:`Skyfall 36B v2 is an enhanced iteration of Mistral Small 2501, specifically fine-tuned for improved creativity, nuanced writing, role-playing, and coherent storytelling.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.55,output:.8},parameters:36,release_date:`2025-03-10`,tagline:`Skyfall 36B v2 is an enhanced iteration of Mistral Small 2501, specifically fine-tuned for improved creativity, nuanced writing, role-playing, and coherent storytelling.`},{id:`thedrummer/unslopnemo-12b`,name:`TheDrummer: UnslopNemo 12B`,created_by:`thedrummer`,source:`official`,last_updated:`2026-03-23`,description:`UnslopNemo v4.1 is the latest addition from the creator of Rocinante, designed for adventure writing and role-play scenarios.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:.4},parameters:12,release_date:`2024-11-08`,tagline:`UnslopNemo v4.1 is the latest addition from the creator of Rocinante, designed for adventure writing and role-play scenarios.`,tools:[`function_calling`]},{id:`tngtech/deepseek-r1t2-chimera`,name:`TNG: DeepSeek R1T2 Chimera`,created_by:`tngtech`,source:`official`,last_updated:`2026-03-23`,family:`deepseek-r1`,description:`DeepSeek-TNG-R1T2-Chimera is the second-generation Chimera model from TNG Tech.`,context_window:163840,max_output_tokens:163840,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.3,output:1.1,cached_input:.15},model_type:`reasoning`,release_date:`2025-07-08`,tagline:`DeepSeek-TNG-R1T2-Chimera is the second-generation Chimera model from TNG Tech.`,tools:[`function_calling`]},{id:`undi95/remm-slerp-l2-13b`,name:`ReMM SLERP 13B`,created_by:`undi95`,source:`official`,last_updated:`2026-03-23`,description:`A recreation trial of the original MythoMax-L2-B13 but with updated models.`,context_window:6144,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.45,output:.65},parameters:13,release_date:`2023-07-22`,tagline:`A recreation trial of the original MythoMax-L2-B13 but with updated models.`},{id:`upstage/solar-pro-3`,name:`Upstage: Solar Pro 3`,created_by:`upstage`,source:`official`,last_updated:`2026-03-23`,description:`Solar Pro 3 is Upstage's powerful Mixture-of-Experts (MoE) language model.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.15,output:.6,cached_input:.015},release_date:`2026-01-27`,tagline:`Solar Pro 3 is Upstage's powerful Mixture-of-Experts (MoE) language model.`,tools:[`function_calling`]},{id:`writer/palmyra-x5`,name:`Writer: Palmyra X5`,created_by:`writer`,source:`official`,last_updated:`2026-03-23`,description:`Palmyra X5 is Writer's most advanced model, purpose-built for building and scaling AI agents across the enterprise.`,context_window:104e4,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},pricing:{input:.6,output:6},model_type:`chat`,release_date:`2026-01-21`,tagline:`Palmyra X5 is Writer's most advanced model, purpose-built for building and scaling AI agents across the enterprise.`},{id:`x-ai/grok-3-beta`,name:`xAI: Grok 3 Beta`,created_by:`xai`,source:`official`,last_updated:`2026-03-23`,family:`grok-3`,description:`Grok 3 is the latest model from xAI.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:3,output:15,cached_input:.75},model_type:`chat`,release_date:`2025-04-09`,tagline:`Grok 3 is the latest model from xAI.`,tools:[`function_calling`]},{id:`x-ai/grok-3-mini-beta`,name:`xAI: Grok 3 Mini Beta`,created_by:`xai`,source:`official`,last_updated:`2026-03-23`,family:`grok-3`,description:`Grok 3 Mini is a lightweight, smaller thinking model.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.3,output:.5,cached_input:.075},model_type:`chat`,release_date:`2025-04-09`,tagline:`Grok 3 Mini is a lightweight, smaller thinking model.`,tools:[`function_calling`]},{id:`x-ai/grok-3-mini`,name:`xAI: Grok 3 Mini`,created_by:`xai`,source:`official`,last_updated:`2026-03-23`,family:`grok-3`,description:`A lightweight model that thinks before responding.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.3,output:.5,cached_input:.075},model_type:`chat`,release_date:`2025-06-10`,tagline:`A lightweight model that thinks before responding.`,tools:[`function_calling`],license:`proprietary`},{id:`x-ai/grok-3`,name:`xAI: Grok 3`,created_by:`xai`,source:`official`,last_updated:`2026-03-23`,family:`grok-3`,description:`Grok 3 is the latest model from xAI.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,streaming:!0},pricing:{input:3,output:15,cached_input:.75},model_type:`chat`,release_date:`2025-06-10`,tagline:`Grok 3 is the latest model from xAI.`,tools:[`function_calling`],license:`proprietary`},{id:`x-ai/grok-4-fast`,name:`xAI: Grok 4 Fast`,created_by:`xai`,source:`official`,last_updated:`2026-03-23`,family:`grok-4`,description:`Grok 4 Fast is xAI's latest multimodal model with SOTA cost-efficiency and a 2M token context window.`,context_window:2e6,max_output_tokens:3e4,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.2,output:.5,cached_input:.05},model_type:`chat`,release_date:`2025-09-19`,tagline:`Grok 4 Fast is xAI's latest multimodal model with SOTA cost-efficiency and a 2M token context window.`,tools:[`function_calling`]},{id:`x-ai/grok-4.1-fast`,name:`xAI: Grok 4.1 Fast`,created_by:`xai`,source:`official`,last_updated:`2026-03-23`,family:`grok-4.1`,description:`Grok 4.1 Fast is xAI's best agentic tool calling model that shines in real-world use cases like customer support and deep research.`,context_window:2e6,max_output_tokens:3e4,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.2,output:.5,cached_input:.05},model_type:`chat`,release_date:`2025-11-19`,tagline:`Grok 4.1 Fast is xAI's best agentic tool calling model that shines in real-world use cases like customer support and deep research.`,tools:[`function_calling`]},{id:`x-ai/grok-4.20-beta`,name:`xAI: Grok 4.20 Beta`,created_by:`xai`,source:`official`,last_updated:`2026-03-23`,family:`grok-4.20`,description:`Grok 4.20 Beta is xAI's newest flagship model with industry-leading speed and agentic tool calling capabilities.`,context_window:2e6,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:2,output:6,cached_input:.2},model_type:`chat`,release_date:`2026-03-12`,tagline:`Grok 4.20 Beta is xAI's newest flagship model with industry-leading speed and agentic tool calling capabilities.`,tools:[`function_calling`]},{id:`x-ai/grok-4.20-multi-agent-beta`,name:`xAI: Grok 4.20 Multi-Agent Beta`,created_by:`xai`,source:`official`,last_updated:`2026-03-23`,family:`grok-4.20`,description:`Grok 4.20 Multi-Agent Beta is a variant of xAI’s Grok 4.20 designed for collaborative, agent-based workflows.`,context_window:2e6,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:2,output:6,cached_input:.2},model_type:`chat`,release_date:`2026-03-12`,tagline:`Grok 4.20 Multi-Agent Beta is a variant of xAI’s Grok 4.20 designed for collaborative, agent-based workflows.`},{id:`x-ai/grok-4`,name:`xAI: Grok 4`,created_by:`xai`,source:`official`,last_updated:`2026-03-23`,family:`grok-4`,description:`Grok 4 is xAI's latest reasoning model with a 256k context window.`,context_window:256e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:3,output:15,cached_input:.75},model_type:`chat`,release_date:`2025-07-09`,tagline:`Grok 4 is xAI's latest reasoning model with a 256k context window.`,tools:[`function_calling`]},{id:`x-ai/grok-code-fast-1`,name:`xAI: Grok Code Fast 1`,created_by:`xai`,source:`official`,last_updated:`2026-03-23`,family:`grok-code`,description:`Grok Code Fast 1 is a speedy and economical reasoning model that excels at agentic coding.`,context_window:256e3,max_output_tokens:1e4,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0,streaming:!0},pricing:{input:.2,output:1.5,cached_input:.02},model_type:`code`,release_date:`2025-08-26`,tagline:`Grok Code Fast 1 is a speedy and economical reasoning model that excels at agentic coding.`,tools:[`function_calling`],license:`proprietary`},{id:`xiaomi/mimo-v2-flash`,name:`Xiaomi: MiMo-V2-Flash`,created_by:`xiaomi`,source:`official`,last_updated:`2026-03-23`,description:`MiMo-V2-Flash is an open-source foundation language model developed by Xiaomi.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.09,output:.29,cached_input:.045},release_date:`2025-12-14`,tagline:`MiMo-V2-Flash is an open-source foundation language model developed by Xiaomi.`,tools:[`function_calling`]},{id:`xiaomi/mimo-v2-omni`,name:`Xiaomi: MiMo-V2-Omni`,created_by:`xiaomi`,source:`official`,last_updated:`2026-03-23`,description:`MiMo-V2-Omni is a frontier omni-modal model that natively processes image, video, and audio inputs within a unified architecture.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`,`audio`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.4,output:2,cached_input:.08},release_date:`2026-03-18`,tagline:`MiMo-V2-Omni is a frontier omni-modal model that natively processes image, video, and audio inputs within a unified architecture.`,tools:[`function_calling`]},{id:`xiaomi/mimo-v2-pro`,name:`Xiaomi: MiMo-V2-Pro`,created_by:`xiaomi`,source:`official`,last_updated:`2026-03-23`,description:`MiMo-V2-Pro is Xiaomi's flagship foundation model, featuring over 1T total parameters and a 1M context length, deeply optimized for agentic scenarios.`,context_window:1048576,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1,output:3,cached_input:.2},release_date:`2026-03-18`,tagline:`MiMo-V2-Pro is Xiaomi's flagship foundation model, featuring over 1T total parameters and a 1M context length, deeply optimized for agentic scenarios.`,tools:[`function_calling`]},{id:`z-ai/glm-4-32b`,name:`Z.ai: GLM 4 32B `,created_by:`z-ai`,source:`official`,last_updated:`2026-03-23`,description:`GLM 4 32B is a cost-effective foundation language model.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.1,output:.1},family:`glm-4`,model_type:`chat`,parameters:32,release_date:`2025-07-24`,tagline:`GLM 4 32B is a cost-effective foundation language model.`,tools:[`function_calling`]},{id:`z-ai/glm-4.5-air:free`,name:`Z.ai: GLM 4.5 Air (free)`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-23`,description:`GLM-4.5-Air is the lightweight variant of our latest flagship model family, also purpose-built for agent-centric applications.`,context_window:131072,max_output_tokens:96e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},family:`glm-4.5`,model_type:`chat`,release_date:`2025-07-25`,tagline:`GLM-4.5-Air is the lightweight variant of our latest flagship model family, also purpose-built for agent-centric applications.`,tools:[`function_calling`]},{id:`z-ai/glm-4.5-air`,name:`Z.ai: GLM 4.5 Air`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-23`,description:`GLM-4.5-Air is the lightweight variant of our latest flagship model family, also purpose-built for agent-centric applications.`,context_window:131072,max_output_tokens:98304,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.13,output:.85,cached_input:.025},family:`glm-4.5`,model_type:`chat`,release_date:`2025-07-25`,tagline:`GLM-4.5-Air is the lightweight variant of our latest flagship model family, also purpose-built for agent-centric applications.`,tools:[`function_calling`]},{id:`z-ai/glm-4.5`,name:`Z.ai: GLM 4.5`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-23`,description:`GLM-4.5 is our latest flagship foundation model, purpose-built for agent-based applications.`,context_window:131072,max_output_tokens:98304,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.6,output:2.2,cached_input:.11},family:`glm-4.5`,model_type:`chat`,release_date:`2025-07-25`,tagline:`GLM-4.5 is our latest flagship foundation model, purpose-built for agent-based applications.`,tools:[`function_calling`]},{id:`z-ai/glm-4.5v`,name:`Z.ai: GLM 4.5V`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-23`,description:`GLM-4.5V is a vision-language foundation model for multimodal agent applications.`,context_window:65536,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.6,output:1.8,cached_input:.11},family:`glm-4.5`,model_type:`chat`,release_date:`2025-08-11`,tagline:`GLM-4.5V is a vision-language foundation model for multimodal agent applications.`,tools:[`function_calling`]},{id:`z-ai/glm-4.6`,name:`Z.ai: GLM 4.6`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-23`,description:`Compared with GLM-4.5, this generation brings several key improvements:
 
-- SoTA understanding of images of various resolution & ratio: Qwen2.5-VL achieves state-of-the-art performanc`,context_window:32768,modalities:{input:[`text`,`image`],output:[`text`]},pricing:{input:.2,output:.2}},{id:`qwen/qwen-max`,name:`Qwen: Qwen-Max `,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen`,description:`Qwen-Max, based on Qwen2.5, provides the best inference performance among [Qwen models](/qwen), especially for complex multi-step tasks.`,context_window:32768,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:1.04,output:4.16,cached_input:.208}},{id:`qwen/qwen-plus-2025-07-28:thinking`,name:`Qwen: Qwen Plus 0728 (thinking)`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen`,description:`Qwen Plus 0728, based on the Qwen3 foundation model, is a 1 million context hybrid reasoning model with a balanced performance, speed, and cost combination.`,context_window:1e6,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.26,output:.78}},{id:`qwen/qwen-plus-2025-07-28`,name:`Qwen: Qwen Plus 0728`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen`,description:`Qwen Plus 0728, based on the Qwen3 foundation model, is a 1 million context hybrid reasoning model with a balanced performance, speed, and cost combination.`,context_window:1e6,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.26,output:.78}},{id:`qwen/qwen-plus`,name:`Qwen: Qwen-Plus`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen`,description:`Qwen-Plus, based on the Qwen2.5 foundation model, is a 131K context model with a balanced performance, speed, and cost combination.`,context_window:1e6,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.26,output:.78,cached_input:.052}},{id:`qwen/qwen-turbo`,name:`Qwen: Qwen-Turbo`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen`,description:`Qwen-Turbo, based on Qwen2.5, is a 1M context model that provides fast speed and low cost, suitable for simple tasks.`,context_window:131072,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.033,output:.13,cached_input:.007}},{id:`qwen/qwen-vl-max`,name:`Qwen: Qwen VL Max`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen`,description:`Qwen VL Max is a visual understanding model with 7500 tokens context length.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.52,output:2.08}},{id:`qwen/qwen-vl-plus`,name:`Qwen: Qwen VL Plus`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen`,description:`Qwen's Enhanced Large Visual Language Model.`,context_window:131072,max_output_tokens:8192,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.137,output:.41,cached_input:.027}},{id:`qwen/qwen2.5-coder-7b-instruct`,name:`Qwen: Qwen2.5 Coder 7B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen2.5`,description:`Qwen2.5-Coder-7B-Instruct is a 7B parameter instruction-tuned language model optimized for code-related tasks such as code generation, reasoning, and bug fixing.`,context_window:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.03,output:.09}},{id:`qwen/qwen2.5-vl-32b-instruct`,name:`Qwen: Qwen2.5 VL 32B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen2.5`,description:`Qwen2.5-VL-32B is a multimodal vision-language model fine-tuned through reinforcement learning for enhanced mathematical reasoning, structured outputs, and visual problem-solving capabilities.`,context_window:128e3,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.2,output:.6}},{id:`qwen/qwen2.5-vl-72b-instruct`,name:`Qwen: Qwen2.5 VL 72B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen2.5`,description:`Qwen2.5-VL is proficient in recognizing common objects such as flowers, birds, fish, and insects.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.8,output:.8}},{id:`qwen/qwen3-14b`,name:`Qwen: Qwen3 14B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-14B is a dense 14.8B parameter causal language model from the Qwen3 series, designed for both complex reasoning and efficient dialogue.`,context_window:40960,max_output_tokens:40960,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.06,output:.24}},{id:`qwen/qwen3-235b-a22b-2507`,name:`Qwen: Qwen3 235B A22B Instruct 2507`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-235B-A22B-Instruct-2507 is a multilingual, instruction-tuned mixture-of-experts language model based on the Qwen3-235B architecture, with 22B active parameters per forward pass.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.071,output:.1}},{id:`qwen/qwen3-235b-a22b-thinking-2507`,name:`Qwen: Qwen3 235B A22B Thinking 2507`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.15,output:1.495}},{id:`qwen/qwen3-235b-a22b`,name:`Qwen: Qwen3 235B A22B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-235B-A22B is a 235B parameter mixture-of-experts (MoE) model developed by Qwen, activating 22B parameters per forward pass.`,context_window:131072,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.455,output:1.82}},{id:`qwen/qwen3-30b-a3b-instruct-2507`,name:`Qwen: Qwen3 30B A3B Instruct 2507`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-30B-A3B-Instruct-2507 is a 30.5B-parameter mixture-of-experts language model from Qwen, with 3.3B active parameters per inference.`,context_window:262144,max_output_tokens:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.09,output:.3}},{id:`qwen/qwen3-30b-a3b-thinking-2507`,name:`Qwen: Qwen3 30B A3B Thinking 2507`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-30B-A3B-Thinking-2507 is a 30B parameter Mixture-of-Experts reasoning model optimized for complex tasks requiring extended multi-step thinking.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.08,output:.4,cached_input:.08}},{id:`qwen/qwen3-30b-a3b`,name:`Qwen: Qwen3 30B A3B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3, the latest generation in the Qwen large language model series, features both dense and mixture-of-experts (MoE) architectures to excel in reasoning, multilingual support, and advanced agent tasks.`,context_window:40960,max_output_tokens:40960,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.08,output:.28}},{id:`qwen/qwen3-32b`,name:`Qwen: Qwen3 32B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-32B is a dense 32.8B parameter causal language model from the Qwen3 series, optimized for both complex reasoning and efficient dialogue.`,context_window:40960,max_output_tokens:40960,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.08,output:.24,cached_input:.04}},{id:`qwen/qwen3-4b:free`,name:`Qwen: Qwen3 4B (free)`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-4B is a 4 billion parameter dense language model from the Qwen3 series, designed to support both general-purpose and reasoning-intensive tasks.`,status:`deprecated`,deprecation_date:`2026-03-29`,context_window:40960,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0}},{id:`qwen/qwen3-8b`,name:`Qwen: Qwen3 8B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-8B is a dense 8.2B parameter causal language model from the Qwen3 series, designed for both reasoning-heavy tasks and efficient dialogue.`,context_window:40960,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.05,output:.4,cached_input:.05}},{id:`qwen/qwen3-coder-30b-a3b-instruct`,name:`Qwen: Qwen3 Coder 30B A3B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-Coder-30B-A3B-Instruct is a 30.5B parameter Mixture-of-Experts (MoE) model with 128 experts (8 active per forward pass), designed for advanced code generation, repository-scale understanding, and agentic tool use.`,context_window:16e4,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.07,output:.27}},{id:`qwen/qwen3-coder-flash`,name:`Qwen: Qwen3 Coder Flash`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3 Coder Flash is Alibaba's fast and cost efficient version of their proprietary Qwen3 Coder Plus.`,context_window:1e6,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.195,output:.975,cached_input:.039}},{id:`qwen/qwen3-coder:free`,name:`Qwen: Qwen3 Coder 480B A35B (free)`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-Coder-480B-A35B-Instruct is a Mixture-of-Experts (MoE) code generation model developed by the Qwen team.`,context_window:262e3,max_output_tokens:262e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0}},{id:`qwen/qwen3-coder-next`,name:`Qwen: Qwen3 Coder Next`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-Coder-Next is an open-weight causal language model optimized for coding agents and local development workflows.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.12,output:.75,cached_input:.06}},{id:`qwen/qwen3-coder-plus`,name:`Qwen: Qwen3 Coder Plus`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3 Coder Plus is Alibaba's proprietary version of the Open Source Qwen3 Coder 480B A35B.`,context_window:1e6,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.65,output:3.25,cached_input:.13}},{id:`qwen/qwen3-coder`,name:`Qwen: Qwen3 Coder 480B A35B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-Coder-480B-A35B-Instruct is a Mixture-of-Experts (MoE) code generation model developed by the Qwen team.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.22,output:1,cached_input:.022}},{id:`qwen/qwen3-max-thinking`,name:`Qwen: Qwen3 Max Thinking`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-Max-Thinking is the flagship reasoning model in the Qwen3 series, designed for high-stakes cognitive tasks that require deep, multi-step reasoning.`,context_window:262144,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.78,output:3.9}},{id:`qwen/qwen3-max`,name:`Qwen: Qwen3 Max`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-Max is an updated release built on the Qwen3 series, offering major improvements in reasoning, instruction following, multilingual support, and long-tail knowledge coverage compared to the January 2025 version.`,context_window:262144,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.78,output:3.9,cached_input:.156}},{id:`qwen/qwen3-next-80b-a3b-instruct:free`,name:`Qwen: Qwen3 Next 80B A3B Instruct (free)`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-Next-80B-A3B-Instruct is an instruction-tuned chat model in the Qwen3-Next series optimized for fast, stable responses without “thinking” traces.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0}},{id:`qwen/qwen3-next-80b-a3b-instruct`,name:`Qwen: Qwen3 Next 80B A3B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-Next-80B-A3B-Instruct is an instruction-tuned chat model in the Qwen3-Next series optimized for fast, stable responses without “thinking” traces.`,context_window:262144,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.09,output:1.1}},{id:`qwen/qwen3-next-80b-a3b-thinking`,name:`Qwen: Qwen3 Next 80B A3B Thinking`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-Next-80B-A3B-Thinking is a reasoning-first chat model in the Qwen3-Next line that outputs structured “thinking” traces by default.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.098,output:.78}},{id:`qwen/qwen3-vl-235b-a22b-instruct`,name:`Qwen: Qwen3 VL 235B A22B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-VL-235B-A22B Instruct is an open-weight multimodal model that unifies strong text generation with visual understanding across images and video.`,context_window:262144,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.2,output:.88,cached_input:.11}},{id:`qwen/qwen3-vl-235b-a22b-thinking`,name:`Qwen: Qwen3 VL 235B A22B Thinking`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-VL-235B-A22B Thinking is a multimodal model that unifies strong text generation with visual understanding across images and video.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.26,output:2.6}},{id:`qwen/qwen3-vl-30b-a3b-instruct`,name:`Qwen: Qwen3 VL 30B A3B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-VL-30B-A3B-Instruct is a multimodal model that unifies strong text generation with visual understanding for images and videos.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.13,output:.52}},{id:`qwen/qwen3-vl-30b-a3b-thinking`,name:`Qwen: Qwen3 VL 30B A3B Thinking`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-VL-30B-A3B-Thinking is a multimodal model that unifies strong text generation with visual understanding for images and videos.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.13,output:1.56}},{id:`qwen/qwen3-vl-32b-instruct`,name:`Qwen: Qwen3 VL 32B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-VL-32B-Instruct is a large-scale multimodal vision-language model designed for high-precision understanding and reasoning across text, images, and video.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.104,output:.416}},{id:`qwen/qwen3-vl-8b-instruct`,name:`Qwen: Qwen3 VL 8B Instruct`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-VL-8B-Instruct is a multimodal vision-language model from the Qwen3-VL series, built for high-fidelity understanding and reasoning across text, images, and video.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.08,output:.5}},{id:`qwen/qwen3-vl-8b-thinking`,name:`Qwen: Qwen3 VL 8B Thinking`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3`,description:`Qwen3-VL-8B-Thinking is the reasoning-optimized variant of the Qwen3-VL-8B multimodal model, designed for advanced visual and textual reasoning across complex scenes, documents, and temporal sequences.`,context_window:131072,max_output_tokens:32768,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.117,output:1.365}},{id:`qwen/qwen3.5-122b-a10b`,name:`Qwen: Qwen3.5-122B-A10B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3.5`,description:`The Qwen3.5 122B-A10B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.26,output:2.08}},{id:`qwen/qwen3.5-27b`,name:`Qwen: Qwen3.5-27B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3.5`,description:`The Qwen3.5 27B native vision-language Dense model incorporates a linear attention mechanism, delivering fast response times while balancing inference speed and performance.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.195,output:1.56}},{id:`qwen/qwen3.5-35b-a3b`,name:`Qwen: Qwen3.5-35B-A3B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3.5`,description:`The Qwen3.5 Series 35B-A3B is a native vision-language model designed with a hybrid architecture that integrates linear attention mechanisms and a sparse mixture-of-experts model, achieving higher inference efficiency.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.163,output:1.3}},{id:`qwen/qwen3.5-397b-a17b`,name:`Qwen: Qwen3.5 397B A17B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3.5`,description:`The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.39,output:2.34}},{id:`qwen/qwen3.5-9b`,name:`Qwen: Qwen3.5-9B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3.5`,description:`Qwen3.5-9B is a multimodal foundation model from the Qwen3.5 family, designed to deliver strong reasoning, coding, and visual understanding in an efficient 9B-parameter architecture.`,context_window:256e3,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.05,output:.15}},{id:`qwen/qwen3.5-flash-02-23`,name:`Qwen: Qwen3.5-Flash`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3.5`,description:`The Qwen3.5 native vision-language Flash models are built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency.`,context_window:1e6,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.065,output:.26}},{id:`qwen/qwen3.5-plus-02-15`,name:`Qwen: Qwen3.5 Plus 2026-02-15`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwen3.5`,description:`The Qwen3.5 native vision-language series Plus models are built on a hybrid architecture that integrates linear attention mechanisms with sparse mixture-of-experts models, achieving higher inference efficiency.`,context_window:1e6,max_output_tokens:65536,modalities:{input:[`text`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.26,output:1.56}},{id:`qwen/qwq-32b`,name:`Qwen: QwQ 32B`,created_by:`qwen`,source:`official`,last_updated:`2026-03-21`,family:`qwq`,description:`QwQ is the reasoning model of the Qwen series.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:.15,output:.58}},{id:`relace/relace-apply-3`,name:`Relace: Relace Apply 3`,created_by:`relace`,source:`official`,last_updated:`2026-03-21`,description:`Relace Apply 3 is a specialized code-patching LLM that merges AI-suggested edits straight into your source files.`,context_window:256e3,max_output_tokens:128e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:.85,output:1.25}},{id:`relace/relace-search`,name:`Relace: Relace Search`,created_by:`relace`,source:`official`,last_updated:`2026-03-21`,description:"The relace-search model uses 4-12 `view_file` and `grep` tools in parallel to explore a codebase and return relevant files to the user request.",context_window:256e3,max_output_tokens:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:1,output:3}},{id:`sao10k/l3-euryale-70b`,name:`Sao10k: Llama 3 Euryale 70B v2.1`,created_by:`sao10k`,source:`official`,last_updated:`2026-03-21`,description:`Euryale 70B v2.1 is a model focused on creative roleplay from [Sao10k](https://ko-fi.com/sao10k).`,context_window:8192,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:1.48,output:1.48}},{id:`sao10k/l3-lunaris-8b`,name:`Sao10K: Llama 3 8B Lunaris`,created_by:`sao10k`,source:`official`,last_updated:`2026-03-21`,description:`Lunaris 8B is a versatile generalist and roleplaying model based on Llama 3.`,context_window:8192,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.04,output:.05}},{id:`sao10k/l3.1-70b-hanami-x1`,name:`Sao10K: Llama 3.1 70B Hanami x1`,created_by:`sao10k`,source:`official`,last_updated:`2026-03-21`,description:`This is [Sao10K](/sao10k)'s experiment over [Euryale v2.2](/sao10k/l3.1-euryale-70b).`,context_window:16e3,modalities:{input:[`text`],output:[`text`]},pricing:{input:3,output:3}},{id:`sao10k/l3.1-euryale-70b`,name:`Sao10K: Llama 3.1 Euryale 70B v2.2`,created_by:`sao10k`,source:`official`,last_updated:`2026-03-21`,description:`Euryale L3.1 70B v2.2 is a model focused on creative roleplay from [Sao10k](https://ko-fi.com/sao10k).`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.85,output:.85}},{id:`sao10k/l3.3-euryale-70b`,name:`Sao10K: Llama 3.3 Euryale 70B`,created_by:`sao10k`,source:`official`,last_updated:`2026-03-21`,description:`Euryale L3.3 70B is a model focused on creative roleplay from [Sao10k](https://ko-fi.com/sao10k).`,context_window:131072,max_output_tokens:16384,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.65,output:.75}},{id:`stepfun/step-3.5-flash:free`,name:`StepFun: Step 3.5 Flash (free)`,created_by:`stepfun`,source:`official`,last_updated:`2026-03-21`,description:`Step 3.5 Flash is StepFun's most capable open-source foundation model.`,context_window:256e3,max_output_tokens:256e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0}},{id:`stepfun/step-3.5-flash`,name:`StepFun: Step 3.5 Flash`,created_by:`stepfun`,source:`official`,last_updated:`2026-03-21`,description:`Step 3.5 Flash is StepFun's most capable open-source foundation model.`,context_window:256e3,max_output_tokens:256e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0},pricing:{input:.1,output:.3,cached_input:.02}},{id:`switchpoint/router`,name:`Switchpoint Router`,created_by:`switchpoint`,source:`official`,last_updated:`2026-03-21`,description:`Switchpoint AI's router instantly analyzes your request and directs it to the optimal AI from an ever-evolving library.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{reasoning:!0},pricing:{input:.85,output:3.4}},{id:`tencent/hunyuan-a13b-instruct`,name:`Tencent: Hunyuan A13B Instruct`,created_by:`tencent`,source:`official`,last_updated:`2026-03-21`,description:`Hunyuan-A13B is a 13B active parameter Mixture-of-Experts (MoE) language model developed by Tencent, with a total parameter count of 80B and support for reasoning via Chain-of-Thought.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:.14,output:.57}},{id:`thedrummer/cydonia-24b-v4.1`,name:`TheDrummer: Cydonia 24B V4.1`,created_by:`thedrummer`,source:`official`,last_updated:`2026-03-21`,description:`Uncensored and creative writing model based on Mistral Small 3.2 24B with good recall, prompt adherence, and intelligence.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},pricing:{input:.3,output:.5}},{id:`thedrummer/rocinante-12b`,name:`TheDrummer: Rocinante 12B`,created_by:`thedrummer`,source:`official`,last_updated:`2026-03-21`,description:`Rocinante 12B is designed for engaging storytelling and rich prose.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.17,output:.43}},{id:`thedrummer/skyfall-36b-v2`,name:`TheDrummer: Skyfall 36B V2`,created_by:`thedrummer`,source:`official`,last_updated:`2026-03-21`,description:`Skyfall 36B v2 is an enhanced iteration of Mistral Small 2501, specifically fine-tuned for improved creativity, nuanced writing, role-playing, and coherent storytelling.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},pricing:{input:.55,output:.8}},{id:`thedrummer/unslopnemo-12b`,name:`TheDrummer: UnslopNemo 12B`,created_by:`thedrummer`,source:`official`,last_updated:`2026-03-21`,description:`UnslopNemo v4.1 is the latest addition from the creator of Rocinante, designed for adventure writing and role-play scenarios.`,context_window:32768,max_output_tokens:32768,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:.4,output:.4}},{id:`tngtech/deepseek-r1t2-chimera`,name:`TNG: DeepSeek R1T2 Chimera`,created_by:`tngtech`,source:`official`,last_updated:`2026-03-21`,family:`deepseek-r1`,description:`DeepSeek-TNG-R1T2-Chimera is the second-generation Chimera model from TNG Tech.`,context_window:163840,max_output_tokens:163840,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.25,output:.85,cached_input:.125}},{id:`undi95/remm-slerp-l2-13b`,name:`ReMM SLERP 13B`,created_by:`undi95`,source:`official`,last_updated:`2026-03-21`,description:`A recreation trial of the original MythoMax-L2-B13 but with updated models.`,context_window:6144,max_output_tokens:4096,modalities:{input:[`text`],output:[`text`]},capabilities:{structured_output:!0},pricing:{input:.45,output:.65}},{id:`upstage/solar-pro-3`,name:`Upstage: Solar Pro 3`,created_by:`upstage`,source:`official`,last_updated:`2026-03-21`,description:`Solar Pro 3 is Upstage's powerful Mixture-of-Experts (MoE) language model.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.15,output:.6,cached_input:.015}},{id:`writer/palmyra-x5`,name:`Writer: Palmyra X5`,created_by:`writer`,source:`official`,last_updated:`2026-03-21`,description:`Palmyra X5 is Writer's most advanced model, purpose-built for building and scaling AI agents across the enterprise.`,context_window:104e4,max_output_tokens:8192,modalities:{input:[`text`],output:[`text`]},pricing:{input:.6,output:6}},{id:`x-ai/grok-3-beta`,name:`xAI: Grok 3 Beta`,created_by:`xai`,source:`official`,last_updated:`2026-03-21`,family:`grok-3`,description:`Grok 3 is the latest model from xAI.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:3,output:15,cached_input:.75}},{id:`x-ai/grok-3-mini-beta`,name:`xAI: Grok 3 Mini Beta`,created_by:`xai`,source:`official`,last_updated:`2026-03-21`,family:`grok-3`,description:`Grok 3 Mini is a lightweight, smaller thinking model.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.3,output:.5,cached_input:.075}},{id:`x-ai/grok-3-mini`,name:`xAI: Grok 3 Mini`,created_by:`xai`,source:`official`,last_updated:`2026-03-21`,family:`grok-3`,description:`A lightweight model that thinks before responding.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.3,output:.5,cached_input:.075}},{id:`x-ai/grok-3`,name:`xAI: Grok 3`,created_by:`xai`,source:`official`,last_updated:`2026-03-21`,family:`grok-3`,description:`Grok 3 is the latest model from xAI.`,context_window:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0},pricing:{input:3,output:15,cached_input:.75}},{id:`x-ai/grok-4-fast`,name:`xAI: Grok 4 Fast`,created_by:`xai`,source:`official`,last_updated:`2026-03-21`,family:`grok-4`,description:`Grok 4 Fast is xAI's latest multimodal model with SOTA cost-efficiency and a 2M token context window.`,context_window:2e6,max_output_tokens:3e4,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.2,output:.5,cached_input:.05}},{id:`x-ai/grok-4.1-fast`,name:`xAI: Grok 4.1 Fast`,created_by:`xai`,source:`official`,last_updated:`2026-03-21`,family:`grok-4.1`,description:`Grok 4.1 Fast is xAI's best agentic tool calling model that shines in real-world use cases like customer support and deep research.`,context_window:2e6,max_output_tokens:3e4,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.2,output:.5,cached_input:.05}},{id:`x-ai/grok-4.20-beta`,name:`xAI: Grok 4.20 Beta`,created_by:`xai`,source:`official`,last_updated:`2026-03-21`,family:`grok-4.20`,description:`Grok 4.20 Beta is xAI's newest flagship model with industry-leading speed and agentic tool calling capabilities.`,context_window:2e6,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:2,output:6,cached_input:.2}},{id:`x-ai/grok-4.20-multi-agent-beta`,name:`xAI: Grok 4.20 Multi-Agent Beta`,created_by:`xai`,source:`official`,last_updated:`2026-03-21`,family:`grok-4.20`,description:`Grok 4.20 Multi-Agent Beta is a variant of xAI’s Grok 4.20 designed for collaborative, agent-based workflows.`,context_window:2e6,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{structured_output:!0,reasoning:!0},pricing:{input:2,output:6,cached_input:.2}},{id:`x-ai/grok-4`,name:`xAI: Grok 4`,created_by:`xai`,source:`official`,last_updated:`2026-03-21`,family:`grok-4`,description:`Grok 4 is xAI's latest reasoning model with a 256k context window.`,context_window:256e3,modalities:{input:[`image`,`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:3,output:15,cached_input:.75}},{id:`x-ai/grok-code-fast-1`,name:`xAI: Grok Code Fast 1`,created_by:`xai`,source:`official`,last_updated:`2026-03-21`,family:`grok-code`,description:`Grok Code Fast 1 is a speedy and economical reasoning model that excels at agentic coding.`,context_window:256e3,max_output_tokens:1e4,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.2,output:1.5,cached_input:.02}},{id:`xiaomi/mimo-v2-flash`,name:`Xiaomi: MiMo-V2-Flash`,created_by:`xiaomi`,source:`official`,last_updated:`2026-03-21`,description:`MiMo-V2-Flash is an open-source foundation language model developed by Xiaomi.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.09,output:.29,cached_input:.045}},{id:`xiaomi/mimo-v2-omni`,name:`Xiaomi: MiMo-V2-Omni`,created_by:`xiaomi`,source:`official`,last_updated:`2026-03-21`,description:`MiMo-V2-Omni is a frontier omni-modal model that natively processes image, video, and audio inputs within a unified architecture.`,context_window:262144,max_output_tokens:65536,modalities:{input:[`text`,`audio`,`image`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.4,output:2,cached_input:.08}},{id:`xiaomi/mimo-v2-pro`,name:`Xiaomi: MiMo-V2-Pro`,created_by:`xiaomi`,source:`official`,last_updated:`2026-03-21`,description:`MiMo-V2-Pro is Xiaomi's flagship foundation model, featuring over 1T total parameters and a 1M context length, deeply optimized for agentic scenarios.`,context_window:1048576,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:1,output:3,cached_input:.2}},{id:`z-ai/glm-4-32b`,name:`Z.ai: GLM 4 32B `,created_by:`z-ai`,source:`official`,last_updated:`2026-03-21`,description:`GLM 4 32B is a cost-effective foundation language model.`,context_window:128e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0},pricing:{input:.1,output:.1}},{id:`z-ai/glm-4.5-air:free`,name:`Z.ai: GLM 4.5 Air (free)`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-21`,description:`GLM-4.5-Air is the lightweight variant of our latest flagship model family, also purpose-built for agent-centric applications.`,context_window:131072,max_output_tokens:96e3,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,reasoning:!0}},{id:`z-ai/glm-4.5-air`,name:`Z.ai: GLM 4.5 Air`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-21`,description:`GLM-4.5-Air is the lightweight variant of our latest flagship model family, also purpose-built for agent-centric applications.`,context_window:131072,max_output_tokens:98304,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.13,output:.85,cached_input:.025}},{id:`z-ai/glm-4.5`,name:`Z.ai: GLM 4.5`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-21`,description:`GLM-4.5 is our latest flagship foundation model, purpose-built for agent-based applications.`,context_window:131072,max_output_tokens:98304,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.6,output:2.2,cached_input:.11}},{id:`z-ai/glm-4.5v`,name:`Z.ai: GLM 4.5V`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-21`,description:`GLM-4.5V is a vision-language foundation model for multimodal agent applications.`,context_window:65536,max_output_tokens:16384,modalities:{input:[`text`,`image`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.6,output:1.8,cached_input:.11}},{id:`z-ai/glm-4.6`,name:`Z.ai: GLM 4.6`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-21`,description:`Compared with GLM-4.5, this generation brings several key improvements:
+Longer context window: The context window has been expanded from 128K to 200K tokens, enabling the model to handle more complex`,context_window:204800,max_output_tokens:204800,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.39,output:1.9},family:`glm-4.6`,model_type:`chat`,release_date:`2025-09-30`,tagline:`Compared with GLM-4.5, this generation brings several key improvements:
 
-Longer context window: The context window has been expanded from 128K to 200K tokens, enabling the model to handle more complex`,context_window:204800,max_output_tokens:204800,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.39,output:1.9}},{id:`z-ai/glm-4.6v`,name:`Z.ai: GLM 4.6V`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-21`,description:`GLM-4.6V is a large multimodal model designed for high-fidelity visual understanding and long-context reasoning across images, documents, and mixed media.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`image`,`text`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.3,output:.9}},{id:`z-ai/glm-4.7-flash`,name:`Z.ai: GLM 4.7 Flash`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-21`,description:`As a 30B-class SOTA model, GLM-4.7-Flash offers a new option that balances performance and efficiency.`,context_window:202752,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.06,output:.4,cached_input:.01}},{id:`z-ai/glm-4.7`,name:`Z.ai: GLM 4.7`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-21`,description:`GLM-4.7 is Z.ai’s latest flagship model, featuring upgrades in two key areas: enhanced programming capabilities and more stable multi-step reasoning/execution.`,context_window:202752,max_output_tokens:65535,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.39,output:1.75,cached_input:.195}},{id:`z-ai/glm-5-turbo`,name:`Z.ai: GLM 5 Turbo`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-21`,description:`GLM-5 Turbo is a new model from Z.ai designed for fast inference and strong performance in agent-driven environments such as OpenClaw scenarios.`,context_window:202752,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.96,output:3.2,cached_input:.192}},{id:`z-ai/glm-5`,name:`Z.ai: GLM 5`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-21`,description:`GLM-5 is Z.ai’s flagship open-source foundation model engineered for complex systems design and long-horizon agent workflows.`,context_window:8e4,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.72,output:2.3}}]},t=e.models.map(t=>({...t,provider:e.id}));exports.models=t,exports.provider=e;
+Longer context window: The context window has been expanded from 128K to 200K tokens, enabling the model to handle more complex`,tools:[`function_calling`]},{id:`z-ai/glm-4.6v`,name:`Z.ai: GLM 4.6V`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-23`,description:`GLM-4.6V is a large multimodal model designed for high-fidelity visual understanding and long-context reasoning across images, documents, and mixed media.`,context_window:131072,max_output_tokens:131072,modalities:{input:[`image`,`text`,`video`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.3,output:.9},family:`glm-4.6`,model_type:`chat`,release_date:`2025-12-08`,tagline:`GLM-4.6V is a large multimodal model designed for high-fidelity visual understanding and long-context reasoning across images, documents, and mixed media.`,tools:[`function_calling`]},{id:`z-ai/glm-4.7-flash`,name:`Z.ai: GLM 4.7 Flash`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-23`,description:`As a 30B-class SOTA model, GLM-4.7-Flash offers a new option that balances performance and efficiency.`,context_window:202752,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.06,output:.4,cached_input:.01},family:`glm-4.7`,model_type:`chat`,release_date:`2026-01-19`,tagline:`As a 30B-class SOTA model, GLM-4.7-Flash offers a new option that balances performance and efficiency.`,tools:[`function_calling`]},{id:`z-ai/glm-4.7`,name:`Z.ai: GLM 4.7`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-23`,description:`GLM-4.7 is Z.ai’s latest flagship model, featuring upgrades in two key areas: enhanced programming capabilities and more stable multi-step reasoning/execution.`,context_window:202752,max_output_tokens:65535,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.39,output:1.75,cached_input:.195},family:`glm-4.7`,model_type:`chat`,release_date:`2025-12-22`,tagline:`GLM-4.7 is Z.ai’s latest flagship model, featuring upgrades in two key areas: enhanced programming capabilities and more stable multi-step reasoning/execution.`,tools:[`function_calling`]},{id:`z-ai/glm-5-turbo`,name:`Z.ai: GLM 5 Turbo`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-23`,description:`GLM-5 Turbo is a new model from Z.ai designed for fast inference and strong performance in agent-driven environments such as OpenClaw scenarios.`,context_window:202752,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.96,output:3.2,cached_input:.192},family:`glm-5`,model_type:`chat`,release_date:`2026-03-15`,tagline:`GLM-5 Turbo is a new model from Z.ai designed for fast inference and strong performance in agent-driven environments such as OpenClaw scenarios.`,tools:[`function_calling`]},{id:`z-ai/glm-5`,name:`Z.ai: GLM 5`,created_by:`z-ai`,source:`official`,last_updated:`2026-03-23`,description:`GLM-5 is Z.ai’s flagship open-source foundation model engineered for complex systems design and long-horizon agent workflows.`,context_window:8e4,max_output_tokens:131072,modalities:{input:[`text`],output:[`text`]},capabilities:{tool_call:!0,structured_output:!0,reasoning:!0},pricing:{input:.72,output:2.3},family:`glm-5`,model_type:`chat`,release_date:`2026-02-11`,tagline:`GLM-5 is Z.ai’s flagship open-source foundation model engineered for complex systems design and long-horizon agent workflows.`,tools:[`function_calling`]}]},t=e.models.map(t=>({...t,provider:e.id}));exports.models=t,exports.provider=e;