crfm-helm 0.5.5__py3-none-any.whl → 0.5.6__py3-none-any.whl

helm/config/model_metadata.yaml

@@ -276,6 +276,14 @@ models:
   # Amazon Nova models
   # References for Amazon Nova models:
   # https://aws.amazon.com/ai/generative-ai/nova/
+  - name: amazon/nova-premier-v1:0
+    display_name: Amazon Nova Premier
+    description: Amazon Nova Premier is the most capable model in the Nova family of foundation models. ([blog](https://aws.amazon.com/blogs/aws/amazon-nova-premier-our-most-capable-model-for-complex-tasks-and-teacher-for-model-distillation/))
+    creator_organization_name: Amazon
+    access: limited
+    release_date: 2025-04-30
+    tags: [NOVA_MODEL_TAG, TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG]
+
   - name: amazon/nova-pro-v1:0
     display_name: Amazon Nova Pro
     description: Amazon Nova Pro Model
@@ -507,6 +515,46 @@ models:
     release_date: 2025-02-24
     tags: [ANTHROPIC_CLAUDE_3_MODEL_TAG, TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  - name: anthropic/claude-3-7-sonnet-20250219-thinking-10k
+    display_name: Claude 3.7 Sonnet (20250219, extended thinking)
+    description: Claude 3.7 Sonnet is a Claude 3 family hybrid reasoning model that can produce near-instant responses or extended, step-by-step thinking that is made visible to the user ([blog](https://www.anthropic.com/news/claude-3-7-sonnet)). Extended thinking is enabled with 10k budget tokens.
+    creator_organization_name: Anthropic
+    access: limited
+    release_date: 2025-02-24
+    tags: [ANTHROPIC_CLAUDE_3_MODEL_TAG, TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: anthropic/claude-sonnet-4-20250514
+    display_name: Claude 4 Sonnet (20250514)
+    description: Claude 4 Sonnet is a hybrid model offering two modes - near-instant responses and extended thinking for deeper reasoning ([blog](https://www.anthropic.com/news/claude-4)).
+    creator_organization_name: Anthropic
+    access: limited
+    release_date: 2025-05-14
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: anthropic/claude-sonnet-4-20250514-thinking-10k
+    display_name: Claude 4 Sonnet (20250514, extended thinking)
+    description: Claude 4 Sonnet is a hybrid model offering two modes - near-instant responses and extended thinking for deeper reasoning ([blog](https://www.anthropic.com/news/claude-4)). Extended thinking is enabled with 10k budget tokens.
+    creator_organization_name: Anthropic
+    access: limited
+    release_date: 2025-05-14
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: anthropic/claude-opus-4-20250514
+    display_name: Claude 4 Opus (20250514)
+    description: Claude 4 Opus is a hybrid model offering two modes - near-instant responses and extended thinking for deeper reasoning ([blog](https://www.anthropic.com/news/claude-4)).
+    creator_organization_name: Anthropic
+    access: limited
+    release_date: 2025-05-14
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: anthropic/claude-opus-4-20250514-thinking-10k
+    display_name: Claude 4 Opus (20250514, extended thinking)
+    description: Claude 4 Opus is a hybrid model offering two modes - near-instant responses and extended thinking for deeper reasoning ([blog](https://www.anthropic.com/news/claude-4)). Extended thinking is enabled with 10k budget tokens.
+    creator_organization_name: Anthropic
+    access: limited
+    release_date: 2025-05-14
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   - name: anthropic/stanford-online-all-v4-s3
     display_name: Anthropic-LM v4-s3 (52B)
     description: A 52B parameter language model, trained using reinforcement learning from human feedback [paper](https://arxiv.org/pdf/2204.05862.pdf).
@@ -868,7 +916,7 @@ models:
     # NOTE: The total size of DeepSeek-R3 model1 on HuggingFace is 685B
     num_parameters: 685000000000
     release_date: 2025-01-20
-    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+    tags: [DEPRECATED_MODEL_TAG, TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
   - name: deepseek-ai/deepseek-r1-hide-reasoning
     display_name: DeepSeek R1 (hide reasoning)
@@ -878,8 +926,35 @@ models:
     # NOTE: The total size of DeepSeek-R3 model1 on HuggingFace is 685B
     num_parameters: 685000000000
     release_date: 2025-01-20
-    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+    tags: [DEPRECATED_MODEL_TAG, TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
  
+  - name: deepseek-ai/deepseek-r1-0528
+    display_name: DeepSeek-R1-0528
+    description: DeepSeek-R1-0528 is a minor version upgrade from DeepSeek R1 that has improved its depth of reasoning and inference capabilities by leveraging increased computational resources and introducing algorithmic optimization mechanisms during post-training. ([paper](https://arxiv.org/abs/2501.12948))
+    creator_organization_name: DeepSeek
+    access: open
+    num_parameters: 685000000000
+    release_date: 2025-05-28
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: deepseek-ai/DeepSeek-R1-Distill-Llama-8B
+    display_name: DeepSeek-R1-Distill-Llama-8b
+    description: DeepSeek-R1-Distill-Llama-8b is a model that is distilled from LLaMA 8B model for the DeepSeek-R1 task.
+    creator_organization_name: DeepSeek
+    access: open
+    num_parameters: 8000000000
+    release_date: 2025-01-20
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: deepseek-ai/deepseek-coder-6.7b-instruct
+    display_name: DeepSeek-Coder-6.7b-Instruct
+    description: DeepSeek-Coder-6.7b-Instruct is a model that is fine-tuned from the LLaMA 6.7B model for the DeepSeek-Coder task.
+    creator_organization_name: DeepSeek
+    access: open
+    num_parameters: 6740000000
+    release_date: 2025-01-20
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   # EleutherAI
   - name: eleutherai/gpt-j-6b # Served by GooseAi, HuggingFace and Together.
     display_name: GPT-J (6B)
@@ -1146,6 +1221,14 @@ models:
     release_date: 2025-02-05
     tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, AUDIO_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  - name: google/gemini-2.0-flash-lite-001
+    display_name: Gemini 2.0 Flash Lite
+    description: Gemini 2.0 Flash Lite ([documentation](https://ai.google.dev/gemini-api/docs/models/gemini))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2025-03-25
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, AUDIO_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   - name: google/gemini-2.0-flash-thinking-exp-01-21
     display_name: Gemini 2.0 Flash Thinking (01-21 preview)
     description: Gemini 2.0 Flash Thinking (01-21 preview) ([documentation](https://cloud.google.com/vertex-ai/generative-ai/docs/thinking))
@@ -1162,6 +1245,70 @@ models:
     release_date: 2025-02-05
     tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, AUDIO_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  - name: google/gemini-2.5-flash-lite-preview-06-17
+    display_name: Gemini 2.5 Flash-Lite (06-17 preview)
+    description: Gemini 2.5 Flash-Lite (06-17 preview) ([blog](https://blog.google/products/gemini/gemini-2-5-model-family-expands/))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2025-06-17
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, AUDIO_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: google/gemini-2.5-flash-preview-04-17
+    display_name: Gemini 2.5 Flash (04-17 preview)
+    description: Gemini 2.5 Flash (04-17 preview) ([documentation](https://ai.google.dev/gemini-api/docs/models/gemini))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2025-04-17
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, AUDIO_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: google/gemini-2.5-flash-preview-05-20
+    display_name: Gemini 2.5 Flash (05-20 preview)
+    description: Gemini 2.5 Flash (05-20 preview) ([documentation](https://ai.google.dev/gemini-api/docs/models/gemini))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2025-04-17
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, AUDIO_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: google/gemini-2.5-flash
+    display_name: Gemini 2.5 Flash
+    description: Gemini 2.5 Flash ([documentation](https://ai.google.dev/gemini-api/docs/models/gemini))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2025-06-17
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, AUDIO_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: google/gemini-2.5-pro-exp-03-25
+    display_name: Gemini 2.5 Pro (03-25 experimental)
+    description: Gemini 2.5 Pro (03-25 experimental) ([documentation](https://ai.google.dev/gemini-api/docs/models/gemini))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2025-03-25
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, AUDIO_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: google/gemini-2.5-pro-preview-03-25
+    display_name: Gemini 2.5 Pro (03-25 preview)
+    description: Gemini 2.5 Pro (03-25 preview) ([documentation](https://ai.google.dev/gemini-api/docs/models/gemini))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2025-04-09  # source: https://cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/2-5-pro
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, AUDIO_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: google/gemini-2.5-pro-preview-05-06
+    display_name: Gemini 2.5 Pro (05-06 preview)
+    description: Gemini 2.5 Pro (05-06 preview) ([documentation](https://ai.google.dev/gemini-api/docs/models/gemini))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2025-05-06  # source: https://cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/2-5-pro
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, AUDIO_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: google/gemini-2.5-pro
+    display_name: Gemini 2.5 Pro
+    description: Gemini 2.5 Pro ([documentation](https://ai.google.dev/gemini-api/docs/models/gemini))
+    creator_organization_name: Google
+    access: limited
+    release_date: 2025-06-17
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, AUDIO_LANGUAGE_MODEL_TAG, GOOGLE_GEMINI_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   - name: google/gemma-2b
     display_name: Gemma (2B)
     description: Gemma is a family of lightweight, open models built from the research and technology that Google used to create the Gemini models. ([model card](https://www.kaggle.com/models/google/gemma), [blog post](https://blog.google/technology/developers/gemma-open-models/))
@@ -1360,6 +1507,60 @@ models:
     release_date: 2023-08-22
     tags: [VISION_LANGUAGE_MODEL_TAG, IDEFICS_MODEL_TAG, IDEFICS_INSTRUCT_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
 
+  - name: huggingface/smollm2-135m
+    display_name: SmolLM2 (135M)
+    description: SmolLM2 is a family of compact language models that are capable of solving a wide range of tasks while being lightweight enough to run on-device. ([paper](https://arxiv.org/abs/2502.02737v1))
+    creator_organization_name: HuggingFace
+    access: open
+    num_parameters: 135000000
+    release_date: 2024-10-31
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: huggingface/smollm2-360m
+    display_name: SmolLM2 (360M)
+    description: SmolLM2 is a family of compact language models that are capable of solving a wide range of tasks while being lightweight enough to run on-device. ([paper](https://arxiv.org/abs/2502.02737v1))
+    creator_organization_name: HuggingFace
+    access: open
+    num_parameters: 362000000
+    release_date: 2024-10-31
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: huggingface/smollm2-1.7b
+    display_name: SmolLM2 (1.7B)
+    description: SmolLM2 is a family of compact language models that are capable of solving a wide range of tasks while being lightweight enough to run on-device. ([paper](https://arxiv.org/abs/2502.02737v1))
+    creator_organization_name: HuggingFace
+    access: open
+    num_parameters: 1710000000
+    release_date: 2024-10-31
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG]
+
+  - name: huggingface/smollm2-135m-instruct
+    display_name: SmolLM2 Instruct (135M)
+    description: SmolLM2 is a family of compact language models that are capable of solving a wide range of tasks while being lightweight enough to run on-device. ([paper](https://arxiv.org/abs/2502.02737v1))
+    creator_organization_name: HuggingFace
+    access: open
+    num_parameters: 135000000
+    release_date: 2024-10-31
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: huggingface/smollm2-360m-instruct
+    display_name: SmolLM2 Instruct (360M)
+    description: SmolLM2 is a family of compact language models that are capable of solving a wide range of tasks while being lightweight enough to run on-device. ([paper](https://arxiv.org/abs/2502.02737v1))
+    creator_organization_name: HuggingFace
+    access: open
+    num_parameters: 362000000
+    release_date: 2024-10-31
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: huggingface/smollm2-1.7b-instruct
+    display_name: SmolLM2 Instruct (1.7B)
+    description: SmolLM2 is a family of compact language models that are capable of solving a wide range of tasks while being lightweight enough to run on-device. ([paper](https://arxiv.org/abs/2502.02737v1))
+    creator_organization_name: HuggingFace
+    access: open
+    num_parameters: 1710000000
+    release_date: 2024-10-31
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   ## Text-to-Image Diffusion Models
   - name: huggingface/dreamlike-diffusion-v1-0
     display_name: Dreamlike Diffusion v1.0 (1B)
@@ -1573,6 +1774,16 @@ models:
     release_date: 2023-06-22
     tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  # Marin Community
+  - name: marin-community/marin-8b-instruct
+    display_name: Marin 8B Instruct
+    description: Marin 8B Instruct is an open-source chatbot trained by fine-tuning LLaMA on user-shared conversations collected from ShareGPT.
+    creator_organization_name: Marin Community
+    access: open
+    num_parameters: 8030000000
+    release_date: 2025-05-15
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   # Meta
   - name: meta/opt-iml-175b # NOT SUPPORTED
     display_name: OPT-IML (175B)
@@ -1875,6 +2086,24 @@ models:
     release_date: 2024-12-06
     tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  - name: meta/llama-4-scout-17b-16e-instruct
+    display_name: Llama 4 Scout (17Bx16E) Instruct
+    description: Llama 4 Scout (17Bx16E) Instruct is part of the Llama 4 collection of models are natively multimodal AI models that enable text and multimodal experiences using a mixture-of-experts architecture. ([blog](https://ai.meta.com/blog/llama-4-multimodal-intelligence/))
+    creator_organization_name: Meta
+    access: open
+    num_parameters: 109000000000
+    release_date: 2025-04-05
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: meta/llama-4-maverick-17b-128e-instruct-fp8
+    display_name: Llama 4 Maverick (17Bx128E) Instruct FP8
+    description: Llama 4 Maverick (17Bx128E) Instruct FP8 is part of the Llama 4 collection of models are natively multimodal AI models that enable text and multimodal experiences using a mixture-of-experts architecture. ([blog](https://ai.meta.com/blog/llama-4-multimodal-intelligence/))
+    creator_organization_name: Meta
+    access: open
+    num_parameters: 402000000000
+    release_date: 2025-04-05
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   - name: meta/llama-3-8b-chat
     display_name: Llama 3 Instruct (8B)
     description: Llama 3 is a family of language models that have been trained on more than 15 trillion tokens, and use Grouped-Query Attention (GQA) for improved inference scalability. It used SFT, rejection sampling, PPO and DPO for post-training. ([paper](https://ai.meta.com/research/publications/the-llama-3-herd-of-models/)
@@ -2150,6 +2379,42 @@ models:
     release_date: 2024-04-17
     tags: [TEXT_MODEL_TAG, FULL_FUNCTIONALITY_TEXT_MODEL_TAG]
 
+  - name: allenai/olmo-2-1124-7b-instruct
+    display_name: OLMo 2 7B Instruct November 2024
+    description: OLMo 2 is a family of 7B and 13B models trained on up to 5T tokens. ([blog](https://allenai.org/blog/olmo2))
+    creator_organization_name: Allen Institute for AI
+    access: open
+    num_parameters: 7300000000
+    release_date: 2024-11-26
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: allenai/olmo-2-1124-13b-instruct
+    display_name: OLMo 2 13B Instruct November 2024
+    description: OLMo 2 is a family of 7B and 13B models trained on up to 5T tokens. ([blog](https://allenai.org/blog/olmo2))
+    creator_organization_name: Allen Institute for AI
+    access: open
+    num_parameters: 13700000000
+    release_date: 2024-11-26
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: allenai/olmo-2-0325-32b-instruct
+    display_name: OLMo 2 32B Instruct March 2025
+    description: OLMo 2 32B Instruct March 2025 is trained up to 6T tokens and post-trained using Tulu 3.1. ([blog](https://allenai.org/blog/olmo2-32B))
+    creator_organization_name: Allen Institute for AI
+    access: open
+    num_parameters: 32200000000
+    release_date: 2025-03-13
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: allenai/olmoe-1b-7b-0125-instruct
+    display_name: OLMoE 1B-7B Instruct January 2025
+    description: OLMoE 1B-7B Instruct January 2025 is a fully open language model leveraging sparse Mixture-of-Experts (MoE). It has 7B parameters but uses only 1B per input token. It was pretrained on 5T tokens. ([blog](https://allenai.org/blog/olmoe-an-open-small-and-state-of-the-art-mixture-of-experts-model-c258432d0514), [paper](https://arxiv.org/abs/2409.02060))
+    creator_organization_name: Allen Institute for AI
+    access: open
+    num_parameters: 32200000000
+    release_date: 2025-03-13
+    tags: [TEXT_MODEL_TAG, PARTIAL_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   # Mistral AI
   - name: mistralai/mistral-7b-v0.1
     display_name: Mistral v0.1 (7B)
@@ -2300,6 +2565,14 @@ models:
     release_date: 2023-12-11
     tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  - name: mistralai/mistral-medium-2505
+    display_name: Mistral Medium 3 (2505)
+    description: Mistral Medium 3 (2505) is a language model that is intended to to deliver state-of-the-art performance at lower cost. ([blog](https://mistral.ai/news/mistral-medium-3))
+    creator_organization_name: Mistral AI
+    access: limited
+    release_date: 2025-05-07
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   - name: mistralai/mistral-large-2402
     display_name: Mistral Large (2402)
     description: Mistral Large is a multilingual model with a 32K tokens context window and function-calling capabilities. ([blog](https://mistral.ai/news/mistral-large/))
@@ -2746,6 +3019,30 @@ models:
     release_date: 2024-07-18
     tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  - name: openai/gpt-4.1-2025-04-14
+    display_name: GPT-4.1 (2025-04-14)
+    description: GPT-4.1 (2025-04-14) is a multimdodal model in the GPT-4.1 family, which outperforms the GPT-4o family, with major gains in coding and instruction following. They also have larger context windows of 1 million tokens and are able to better use that context with improved long-context comprehension. ([blog](https://openai.com/index/gpt-4-1/))
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-04-14
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: openai/gpt-4.1-mini-2025-04-14
+    display_name: GPT-4.1 mini (2025-04-14)
+    description: GPT-4.1 mini (2025-04-14) is a multimdodal model in the GPT-4.1 family, which outperforms the GPT-4o family, with major gains in coding and instruction following. They also have larger context windows of 1 million tokens and are able to better use that context with improved long-context comprehension. ([blog](https://openai.com/index/gpt-4-1/))
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-04-14
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: openai/gpt-4.1-nano-2025-04-14
+    display_name: GPT-4.1 nano (2025-04-14)
+    description: GPT-4.1 nano (2025-04-14) is a multimdodal model in the GPT-4.1 family, which outperforms the GPT-4o family, with major gains in coding and instruction following. They also have larger context windows of 1 million tokens and are able to better use that context with improved long-context comprehension. ([blog](https://openai.com/index/gpt-4-1/))
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-04-14
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   - name: openai/whisper-1_gpt-4o-2024-11-20
     display_name: Whisper-1 + GPT-4o (2024-11-20)
     description: Transcribes the text with Whisper-1 and then uses GPT-4o to generate a response.
@@ -2754,6 +3051,22 @@ models:
     release_date: 2024-11-20
     tags: [AUDIO_LANGUAGE_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG]
 
+  - name: openai/gpt-4o-transcribe_gpt-4o-2024-11-20
+    display_name: GPT-4o Transcribe + GPT-4o (2024-11-20)
+    description: Transcribes the text with GPT-4o Transcribe and then uses GPT-4o to generate a response.
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-03-20
+    tags: [AUDIO_LANGUAGE_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG]
+
+  - name: openai/gpt-4o-mini-transcribe_gpt-4o-2024-11-20
+    display_name: GPT-4o mini Transcribe + GPT-4o (2024-11-20)
+    description: Transcribes the text with GPT-4o mini Transcribe and then uses GPT-4o to generate a response.
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-03-20
+    tags: [AUDIO_LANGUAGE_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG]
+
   - name: openai/gpt-4o-audio-preview-2024-10-01
     display_name: GPT-4o Audio (Preview 2024-10-01)
     description: GPT-4o Audio (Preview 2024-10-01) is a preview model that allows using use audio inputs to prompt the model ([documentation](https://platform.openai.com/docs/guides/audio)).
@@ -2807,6 +3120,30 @@ models:
     tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, OPENAI_CHATGPT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
   ## o1 Models
+  - name: openai/o1-pro-2025-03-19
+    display_name: o1 pro (2025-03-19)
+    description: o1 is a new large language model trained with reinforcement learning to perform complex reasoning. ([model card](https://openai.com/index/openai-o1-system-card/), [blog post](https://openai.com/index/learning-to-reason-with-llms/))
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-03-19
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: openai/o1-pro-2025-03-19-low-reasoning-effort
+    display_name: o1 pro (2025-03-19, low reasoning effort)
+    description: o1 is a new large language model trained with reinforcement learning to perform complex reasoning. ([model card](https://openai.com/index/openai-o1-system-card/), [blog post](https://openai.com/index/learning-to-reason-with-llms/)) The requests' reasoning effort parameter in is set to low.
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-03-19
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: openai/o1-pro-2025-03-19-high-reasoning-effort
+    display_name: o1 pro (2025-03-19, high reasoning effort)
+    description: o1 is a new large language model trained with reinforcement learning to perform complex reasoning. ([model card](https://openai.com/index/openai-o1-system-card/), [blog post](https://openai.com/index/learning-to-reason-with-llms/)) The requests' reasoning effort parameter in is set to high.
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-03-19
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   - name: openai/o1-2024-12-17
     display_name: o1 (2024-12-17)
     description: o1 is a new large language model trained with reinforcement learning to perform complex reasoning. ([model card](https://openai.com/index/openai-o1-system-card/), [blog post](https://openai.com/index/learning-to-reason-with-llms/))
@@ -2871,6 +3208,54 @@ models:
     release_date: 2025-01-31
     tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  - name: openai/o3-2025-04-16
+    display_name: o3 (2025-04-16)
+    description: o3 is a reasoning model for math, science, coding, and visual reasoning tasks. ([blog post](https://openai.com/index/introducing-o3-and-o4-mini/))
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-04-16
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: openai/o3-2025-04-16-low-reasoning-effort
+    display_name: o3 (2025-04-16, low reasoning effort)
+    description: o3 is a reasoning model for math, science, coding, and visual reasoning tasks. ([blog post](https://openai.com/index/introducing-o3-and-o4-mini/))
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-04-16
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: openai/o3-2025-04-16-high-reasoning-effort
+    display_name: o3 (2025-04-16, high reasoning effort)
+    description: o3 is a reasoning model for math, science, coding, and visual reasoning tasks. ([blog post](https://openai.com/index/introducing-o3-and-o4-mini/))
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-04-16
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: openai/o4-mini-2025-04-16
+    display_name: o4-mini (2025-04-16)
+    description: o4-mini is an o-series model optimized for fast, effective reasoning with exceptionally efficient performance in coding and visual tasks. ([blog post](https://openai.com/index/introducing-o3-and-o4-mini/))
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-04-16
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: openai/o4-mini-2025-04-16-low-reasoning-effort
+    display_name: o4-mini (2025-04-16, low reasoning effort)
+    description: o4-mini is an o-series model optimized for fast, effective reasoning with exceptionally efficient performance in coding and visual tasks. ([blog post](https://openai.com/index/introducing-o3-and-o4-mini/))
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-04-16
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: openai/o4-mini-2025-04-16-high-reasoning-effort
+    display_name: o4-mini (2025-04-16, high reasoning effort)
+    description: o4-mini is an o-series model optimized for fast, effective reasoning with exceptionally efficient performance in coding and visual tasks. ([blog post](https://openai.com/index/introducing-o3-and-o4-mini/))
+    creator_organization_name: OpenAI
+    access: limited
+    release_date: 2025-04-16
+    tags: [TEXT_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   ## Codex Models
   # DEPRECATED: Codex models have been shut down on March 23 2023.
 
@@ -3139,6 +3524,14 @@ models:
     release_date: 2024-09-19
     tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  - name: qwen/qwen3-235b-a22b-fp8-tput
+    display_name: Qwen3 235B A22B FP8 Throughput
+    description: Qwen3 235B A22B FP8 Throughput is a hybrid instruct and reasoning mixture-of-experts model ([blog](https://qwenlm.github.io/blog/qwen3/)).
+    creator_organization_name: Qwen
+    access: open
+    release_date: 2025-04-29
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   - name: qwen/qwq-32b-preview
     display_name: QwQ (32B Preview)
     description: QwQ-32B-Preview is an experimental research model developed by the Qwen Team, focused on advancing AI reasoning capabilities. ([blog post](https://qwenlm.github.io/blog/qwq-32b-preview/)).
@@ -3180,6 +3573,38 @@ models:
     release_date: 2024-08-29
     tags: [VISION_LANGUAGE_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
 
+  - name: qwen/qwen2.5-vl-3b-instruct
+    display_name: Qwen2.5-VL Instruct (3B)
+    description: The second generation of Qwen2.5-VL models ([blog](https://qwenlm.github.io/blog/qwen2.5-vl/)).
+    creator_organization_name: Alibaba Group
+    access: open
+    release_date: 2025-01-26
+    tags: [VISION_LANGUAGE_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
+
+  - name: qwen/qwen2.5-vl-7b-instruct
+    display_name: Qwen2.5-VL Instruct (7B)
+    description: The second generation of Qwen2.5-VL models ([blog](https://qwenlm.github.io/blog/qwen2.5-vl/)).
+    creator_organization_name: Alibaba Group
+    access: open
+    release_date: 2025-01-26
+    tags: [VISION_LANGUAGE_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
+
+  - name: qwen/qwen2.5-vl-32b-instruct
+    display_name: Qwen2.5-VL Instruct (32B)
+    description: The second generation of Qwen2.5-VL models ([blog](https://qwenlm.github.io/blog/qwen2.5-vl/)).
+    creator_organization_name: Alibaba Group
+    access: open
+    release_date: 2025-01-26
+    tags: [VISION_LANGUAGE_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
+
+  - name: qwen/qwen2.5-vl-72b-instruct
+    display_name: Qwen2.5-VL Instruct (72B)
+    description: The second generation of Qwen2.5-VL models ([blog](https://qwenlm.github.io/blog/qwen2.5-vl/)).
+    creator_organization_name: Alibaba Group
+    access: open
+    release_date: 2025-01-26
+    tags: [VISION_LANGUAGE_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
+
   - name: qwen/qwen-audio-chat
     display_name: Qwen-Audio Chat
     description: Auditory multimodal version of the Qwen large language model series ([paper](https://arxiv.org/abs/2311.07919)).
@@ -3196,6 +3621,14 @@ models:
     release_date: 2024-07-15
     tags: [AUDIO_LANGUAGE_MODEL_TAG]
 
+  - name: qwen/qwen2.5-omni-7b
+    display_name: Qwen2.5-Omni (7B)
+    description: The new flagship end-to-end multimodal model in the Qwen series that can process inputs including text, images, audio, and video ([paper](https://arxiv.org/abs/2503.20215)).
+    creator_organization_name: Alibaba Cloud
+    access: open
+    release_date: 2025-03-27
+    tags: [AUDIO_LANGUAGE_MODEL_TAG, VISION_LANGUAGE_MODEL_TAG, FULL_FUNCTIONALITY_VLM_TAG]
+
   # SAIL (Sea AI Lab)
   - name: sail/sailor-7b
     display_name: Sailor (7B)
@@ -3670,6 +4103,14 @@ models:
     release_date: 2024-09-12
     tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
+  - name: writer/palmyra-x5
+    display_name: Palmyra X5
+    description: Palmyra X5 is a language model for enterprise that uses a Mixture of Experts (MoE) architecture and a hybrid attention mechanism that blends linear and softmax attention. ([blog](https://writer.com/engineering/long-context-palmyra-x5/))
+    creator_organization_name: Writer
+    access: limited
+    release_date: 2024-04-28
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   - name: writer/palmyra-med-32k
     display_name: Palmyra-Med 32K (70B)
     description: Palmyra-Med 32K (70B) is a model finetuned from Palmyra-X-003 intended for medical applications.
@@ -3680,11 +4121,10 @@ models:
     tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
   - name: writer/palmyra-med
-    display_name: Palmyra-Med (70B)
-    description: Palmyra-Med (70B) is a model finetuned from Palmyra-X-003 intended for medical applications.
+    display_name: Palmyra Med
+    description: Palmyra Med is a model intended for medical applications.
     creator_organization_name: Writer
     access: open
-    num_parameters: 70600000000
     release_date: 2024-07-31
     tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
@@ -3707,12 +4147,20 @@ models:
 
   # xAI
 
-  - name: xai/grok-beta
-    display_name: Grok Beta
-    description: Grok Beta is a model from xAI.
+  - name: xai/grok-3-beta
+    display_name: Grok 3 Beta
+    description: Grok 3 Beta is a model trained on xAI's Colossus supercluster with significant improvements in reasoning, mathematics, coding, world knowledge, and instruction-following tasks. ([blog](https://x.ai/news/grok-3))
     creator_organization_name: xAI
-    access: closed
-    release_date: 2024-08-13
+    access: limited
+    release_date: 2025-04-03  # https://docs.x.ai/docs/release-notes#april-2025
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: xai/grok-3-mini-beta
+    display_name: Grok 3 mini Beta
+    description: Grok 3 mini Beta is a model trained on xAI's Colossus supercluster with significant improvements in reasoning, mathematics, coding, world knowledge, and instruction-following tasks. ([blog](https://x.ai/news/grok-3))
+    creator_organization_name: xAI
+    access: limited
+    release_date: 2025-04-03  # https://docs.x.ai/docs/release-notes#april-2025
     tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
   # Yandex
@@ -3807,6 +4255,17 @@ models:
     release_date: 2024-09-10
     tags: [AUDIO_LANGUAGE_MODEL_TAG]
 
+
+# Maritaca AI
+  - name: maritaca-ai/sabia-7b
+    display_name: Sabia 7B
+    description: Sabia 7B
+    creator_organization_name: MARITACA-AI
+    access: open
+    num_parameters: 6740000000
+    release_date: 2023-11-08
+    tags: [TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
 # Granite - IBM
 # https://www.ibm.com/granite
 # https://github.com/ibm-granite/granite-3.0-language-models
@@ -3882,17 +4341,7 @@ models:
     num_parameters: 1380000000
     release: 2024-10-21
     tags: [TEXT_MODEL_TAG]
-
-  - name: maritaca-ai/sabia-7b
-    display_name: Sabia 7B
-    description: Sabia 7B
-    creator_organization_name: MARITACA-AI
-    access: open
-    num_parameters: 6740000000
-    release_date: 2023-11-08
-    tags: [TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
    
-# Granite-3.1-8b-base
   - name: ibm-granite/granite-3.1-8b-base
     display_name: Granite 3.1 - 8B - Base
     description: Granite-3.1-8B-Base extends the context length of Granite-3.0-8B-Base from 4K to 128K using a progressive training strategy by increasing the supported context length in increments while adjusting RoPE theta until the model has successfully adapted to desired length of 128K.
@@ -3902,7 +4351,6 @@ models:
     release_date: 2024-12-18
     tags: [TEXT_MODEL_TAG]
 
-# Granite-3.1-8b-instruct
   - name: ibm-granite/granite-3.1-8b-instruct
     display_name: Granite 3.1 - 8B - Instruct
     description: Granite-3.1-8B-Instruct is a 8B parameter long-context instruct model finetuned from Granite-3.1-8B-Base using a combination of open source instruction datasets with permissive license and internally collected synthetic datasets tailored for solving long context problems.
@@ -3912,7 +4360,6 @@ models:
     release_date: 2024-12-18
     tags: [TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
-# Granite-3.1-2b-instruct
   - name: ibm-granite/granite-3.1-2b-instruct
     display_name: Granite 3.1 - 2B - Instruct
     description: Granite-3.1-2B-Instruct is a 2B parameter long-context instruct model finetuned from Granite-3.1-2B-Base using a combination of open source instruction datasets with permissive license and internally collected synthetic datasets tailored for solving long context problems.
@@ -3922,7 +4369,6 @@ models:
     release_date: 2024-12-18
     tags: [TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
-# Granite-3.1-2b-base
   - name: ibm-granite/granite-3.1-2b-base
     display_name: Granite 3.1 - 2B - Base
     description: Granite-3.1-2B-Base extends the context length of Granite-3.0-2B-Base from 4K to 128K using a progressive training strategy by increasing the supported context length in increments while adjusting RoPE theta until the model has successfully adapted to desired length of 128K.
@@ -3932,7 +4378,6 @@ models:
     release_date: 2024-12-18
     tags: [TEXT_MODEL_TAG]
 
-# Granite-3.1-3b-a800m-instruct
   - name: ibm-granite/granite-3.1-3b-a800m-instruct
     display_name: Granite 3.1 - 3B - A800M - Instruct
     description: Granite-3.1-3B-A800M-Instruct is a 3B parameter long-context instruct model finetuned from Granite-3.1-3B-A800M-Base using a combination of open source instruction datasets with permissive license and internally collected synthetic datasets tailored for solving long context problems.
@@ -3942,7 +4387,6 @@ models:
     release_date: 2024-12-18
     tags: [TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
-# Granite-3.1-3b-a800m-base
   - name: ibm-granite/granite-3.1-3b-a800m-base
     display_name: Granite 3.1 - 3B - A800M - Base
     description: Granite-3.1-3B-A800M-Base extends the context length of Granite-3.0-3B-A800M-Base from 4K to 128K using a progressive training strategy by increasing the supported context length in increments while adjusting RoPE theta until the model has successfully adapted to desired length of 128K.
@@ -3952,7 +4396,6 @@ models:
     release_date: 2024-12-18
     tags: [TEXT_MODEL_TAG]
 
-# Granite-3.1-1b-a400m-instruct
   - name: ibm-granite/granite-3.1-1b-a400m-instruct
     display_name: Granite 3.1 - 1B - A400M - Instruct
     description: Granite-3.1-1B-A400M-Instruct is a 8B parameter long-context instruct model finetuned from Granite-3.1-1B-A400M-Base using a combination of open source instruction datasets with permissive license and internally collected synthetic datasets tailored for solving long context problems.
@@ -3962,7 +4405,6 @@ models:
     release_date: 2024-12-18
     tags: [TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
 
-# Granite-3.1-1b-a400m-base
   - name: ibm-granite/granite-3.1-1b-a400m-base
     display_name: Granite 3.1 - 1B - A400M - Base
     description: Granite-3.1-1B-A400M-Base extends the context length of Granite-3.0-1B-A400M-Base from 4K to 128K using a progressive training strategy by increasing the supported context length in increments while adjusting RoPE theta until the model has successfully adapted to desired length of 128K.
@@ -3972,27 +4414,6 @@ models:
     release_date: 2024-12-18
     tags: [TEXT_MODEL_TAG]
 
-# DeepSeek-R1-Distill-Llama-3.1-8b
-  - name: deepseek-ai/DeepSeek-R1-Distill-Llama-8B
-    display_name: DeepSeek-R1-Distill-Llama-8b
-    description: DeepSeek-R1-Distill-Llama-8b is a model that is distilled from LLaMA 8B model for the DeepSeek-R1 task.
-    creator_organization_name: DeepSeek
-    access: open
-    num_parameters: 8000000000
-    release_date: 2025-01-20
-    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
-
-# deepseek-ai/deepseek-coder-6.7b-instruct
-  - name: deepseek-ai/deepseek-coder-6.7b-instruct
-    display_name: DeepSeek-Coder-6.7b-Instruct
-    description: DeepSeek-Coder-6.7b-Instruct is a model that is fine-tuned from the LLaMA 6.7B model for the DeepSeek-Coder task.
-    creator_organization_name: DeepSeek
-    access: open
-    num_parameters: 6740000000
-    release_date: 2025-01-20
-    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
-
-# WatsonX - IBM
   - name: ibm/granite-13b-instruct-v2
     display_name: Granite 13b instruct v2
     description: Granite Base (13B) Instruct V2.0 is a large decoder-only transformer model.The following features were used in the design of the model Decoder-only model
@@ -4039,11 +4460,6 @@ models:
     release: 2024-6-18
     tags: [ TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG ]
 
-
-
-
-
-
   - name: ibm/granite-3.1-8b-instruct
     display_name: Granite 3.1 - 8B - Instruct
     description: Granite-3.1-8B-Instruct is a 8B parameter long-context instruct model finetuned from Granite-3.1-8B-Base using a combination of open source instruction datasets with permissive license and internally collected synthetic datasets tailored for solving long context problems.
@@ -4062,6 +4478,15 @@ models:
     release_date: 2024-12-18
     tags: [ TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG ]
 
+  - name: ibm/granite-3.3-8b-instruct
+    display_name: Granite 3.3 8B Instruct
+    description: Granite 3.3 8B Instruct is a 8-billion parameter 128K context length language model fine-tuned for improved reasoning and instruction-following capabilities. ([model card](https://huggingface.co/ibm-granite/granite-3.3-8b-instruct))
+    creator_organization_name: IBM
+    access: open
+    num_parameters: 8170000000
+    release_date: 2025-04-16
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, ABLATION_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
   - name: mistralai/mixtral-8x7b-instruct-v0:1
     display_name: Mixtral 8x7B Instruct on IBM WatsonX
     description: A 7B sparse Mixture-of-Experts model with stronger capabilities than Mistral 7B. Uses 12B active parameters out of 45B total. Supports multiple languages, code and 32k context window.
@@ -4069,3 +4494,192 @@ models:
     access: limited
     release_date: 2023-12-11
     tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, ABLATION_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: ura-hcmut/ura-llama-2.1-8b
+    display_name: URA-Llama 2.1 (8B)
+    description: URA-Llama 2.1 (8B) is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: URA
+    access: open
+    num_parameters: 8000000000
+    release_date: 2024-08-04
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: ura-hcmut/ura-llama-2-8b
+    display_name: URA-Llama 2 (8B)
+    description: URA-Llama 2 (8B) is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: URA
+    access: open
+    num_parameters: 8000000000
+    release_date: 2024-08-04
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: ura-hcmut/ura-llama-7b
+    display_name: URA-Llama 7B (7B)
+    description: URA-Llama 7B (7B) is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: URA
+    access: open
+    num_parameters: 7000000000
+    release_date: 2023-10-10
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: ura-hcmut/ura-llama-13b
+    display_name: URA-Llama 13B (13B)
+    description: URA-Llama 13B (13B) is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: URA
+    access: open
+    num_parameters: 13000000000
+    release_date: 2023-10-10
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: ura-hcmut/ura-llama-70b
+    display_name: URA-Llama 70B (70B)
+    description: URA-Llama 70B (70B) is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: URA
+    access: open
+    num_parameters: 70000000000
+    release_date: 2023-10-10
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: ura-hcmut/GemSUra-7B
+    display_name: GemSUra 7B
+    description: GemSUra 7B is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: URA
+    access: open
+    num_parameters: 7000000000
+    release_date: 2024-03-12
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: ura-hcmut/GemSUra-2B
+    display_name: GemSUra 2B
+    description: GemSUra 2B is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: URA
+    access: open
+    num_parameters: 2000000000
+    release_date: 2024-03-12
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: ura-hcmut/MixSUra
+    display_name: MixSUra
+    description: MixSUra is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text. It is a mixture of experts model with 8 active experts.
+    creator_organization_name: URA
+    access: open
+    num_parameters: 46700000000
+    release_date: 2024-03-12
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: vilm/vinallama-7b-chat
+    display_name: VinaLLaMa
+    description: VinaLLaMa is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: ViLM
+    access: open
+    num_parameters: 7000000000
+    release_date: 2024-03-12
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: vilm/vinallama-2.7b-chat
+    display_name: VinaLLaMa 2.7B
+    description: VinaLLaMa 2.7B is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: ViLM
+    access: open
+    num_parameters: 2700000000
+    release_date: 2024-03-12
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: vilm/vietcuna-7b-v3
+    display_name: VietCuna 7B (v3)
+    description: VietCuna 7B is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: ViLM
+    access: open
+    num_parameters: 7000000000
+    release_date: 2023-08-07
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: vilm/vietcuna-3b-v2
+    display_name: VietCuna 3B (v2)
+    description: VietCuna 3B is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: ViLM
+    access: open
+    num_parameters: 3000000000
+    release_date: 2023-07-15
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: vilm/Quyen-v0.1
+    display_name: Quyen (v0.1)
+    description: Quyen is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: ViLM
+    access: open
+    num_parameters: 4000000000
+    release_date: 2024-02-26
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: vilm/Quyen-Plus-v0.1
+    display_name: Quyen Plus (v0.1)
+    description: Quyen Plus is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: ViLM
+    access: open
+    num_parameters: 7000000000
+    release_date: 2024-02-26
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: vilm/Quyen-Pro-v0.1
+    display_name: Quyen Pro (v0.1)
+    description: Quyen Pro is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: ViLM
+    access: open
+    num_parameters: 14000000000
+    release_date: 2024-02-26
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: vilm/Quyen-Pro-Max-v0.1
+    display_name: Quyen Pro Max (v0.1)
+    description: Quyen Pro Max is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: ViLM
+    access: open
+    num_parameters: 72000000000
+    release_date: 2024-02-26
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: vilm/Quyen-Mini-v0.1
+    display_name: Quyen Mini (v0.1)
+    description: Quyen Mini is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: ViLM
+    access: open
+    num_parameters: 1800000000
+    release_date: 2024-02-26
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: vilm/Quyen-SE-v0.1
+    display_name: Quyen SE (v0.1)
+    description: Quyen SE is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: ViLM
+    access: open
+    num_parameters: 500000000
+    release_date: 2024-02-26
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: Viet-Mistral/Vistral-7B-Chat
+    display_name: Vistral 7B Chat
+    description: Vistral 7B Chat is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: Viet-Mistral
+    access: open
+    num_parameters: 7000000000
+    release_date: 2024-02-28
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: vinai/PhoGPT-7B5-Instruct
+    display_name: PhoGPT 7B5 Instruct
+    description: PhoGPT 7B5 Instruct is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: VinAI
+    access: open
+    num_parameters: 7500000000
+    release_date: 2024-02-19
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]
+
+  - name: vinai/PhoGPT-4B-Chat
+    display_name: PhoGPT 4B Chat
+    description: PhoGPT 4B Chat is a model trained on a large corpus of Vietnamese text data, including books, articles, and websites. It is designed to understand and generate Vietnamese text.
+    creator_organization_name: VinAI
+    access: open
+    num_parameters: 4000000000
+    release_date: 2024-04-02
+    tags: [TEXT_MODEL_TAG, LIMITED_FUNCTIONALITY_TEXT_MODEL_TAG, INSTRUCTION_FOLLOWING_MODEL_TAG]