@just-every/ensemble 0.2.260 → 0.2.262

package/README.md

@@ -40,7 +40,7 @@ See the [demo README](demo/README.md) for detailed information about each demo.
 - OpenAI: Added GPT-5.2 (base + chat-latest + pro) and refreshed GPT-5.1/GPT-5/Codex pricing
 - Anthropic: Claude 4.5 (Sonnet/Haiku, incl. 1M context) and Claude Opus 4.1
 - Google: Gemini 3 (Pro/Flash/Ultra) and refreshed Gemini 2.5 pricing incl. image/TTS/native-audio
-- xAI: Grok 4.1 Fast and Grok 4 Fast with tiered pricing; updated Grok 4/3/mini variants plus Grok Imagine image generation/editing support (`grok-imagine-image`, `grok-imagine-image-pro`)
+- xAI: Grok 4.5 with current `grok-build-latest` aliasing and reasoning controls, Grok 4.1 Fast / Grok 4 Fast variants, and Grok Imagine image generation/editing support (`grok-imagine-image`, `grok-imagine-image-pro`)
 
 \*Codex-Max pricing reflects current published rates and may change if OpenAI updates pricing.
 

package/dist/cjs/data/model_data.cjs

@@ -8,8 +8,8 @@ exports.MODEL_CLASSES = {
         models: [
             'gpt-5.5',
             'gemini-3-flash-preview',
-            'claude-sonnet-4-6',
-            'grok-4.3',
+            'claude-sonnet-5',
+            'grok-4.5',
         ],
         random: true,
     },
@@ -26,8 +26,8 @@ exports.MODEL_CLASSES = {
         models: [
             'gpt-5.5',
             'gemini-2.5-pro',
-            'claude-sonnet-4-6',
-            'grok-4.3',
+            'claude-sonnet-5',
+            'grok-4.5',
         ],
         random: true,
     },
@@ -36,7 +36,7 @@ exports.MODEL_CLASSES = {
             'gpt-5.5-pro',
             'gemini-3.1-pro-preview',
             'claude-fable-5',
-            'grok-4.3',
+            'grok-4.5',
         ],
         random: true,
     },
@@ -44,7 +44,7 @@ exports.MODEL_CLASSES = {
         models: [
             'gpt-5.4-mini',
             'gemini-3-flash-preview',
-            'claude-sonnet-4-6',
+            'claude-sonnet-5',
             'grok-3-mini',
         ],
         random: true,
@@ -53,8 +53,8 @@ exports.MODEL_CLASSES = {
         models: [
             'gpt-5.5',
             'gemini-3.1-pro-preview',
-            'claude-sonnet-4-6',
-            'grok-4.3',
+            'claude-sonnet-5',
+            'grok-4.5',
         ],
         random: true,
     },
@@ -63,7 +63,7 @@ exports.MODEL_CLASSES = {
             'gpt-5.5',
             'gemini-3.1-pro-preview',
             'claude-fable-5',
-            'grok-4.3',
+            'grok-4.5',
         ],
         random: true,
     },
@@ -72,7 +72,7 @@ exports.MODEL_CLASSES = {
             'gpt-5.3-codex',
             'gemini-3.1-pro-preview',
             'claude-fable-5',
-            'grok-4.3',
+            'grok-4.5',
             'qwen3-coder',
         ],
         random: true,
@@ -81,8 +81,8 @@ exports.MODEL_CLASSES = {
         models: [
             'gpt-5.5',
             'gemini-3-flash-preview',
-            'claude-sonnet-4-6',
-            'grok-4.3',
+            'claude-sonnet-5',
+            'grok-4.5',
         ],
         random: true,
     },
@@ -100,7 +100,7 @@ exports.MODEL_CLASSES = {
             'gpt-5.5',
             'gemini-3.1-pro-preview',
             'claude-fable-5',
-            'grok-4.3',
+            'grok-4.5',
         ],
         random: true,
     },
@@ -157,7 +157,7 @@ exports.MODEL_CLASSES = {
         models: [
             'tts-1-hd',
             'eleven_multilingual_v2',
-            'gemini-2.5-pro-preview-tts',
+            'gemini-3.1-flash-tts-preview',
         ],
         description: 'Text-to-Speech models for voice generation',
     },
@@ -166,7 +166,7 @@ exports.MODEL_CLASSES = {
             'gpt-realtime-whisper',
             'gpt-4o-transcribe',
             'u3-rt-pro',
-            'gemini-2.5-flash-native-audio-preview-12-2025',
+            'gemini-3.1-flash-live-preview',
         ],
         description: 'Speech-to-Text models for audio transcription with real-time streaming',
     },
@@ -498,6 +498,27 @@ exports.MODEL_REGISTRY = [
         },
         description: 'Qwen3.5-397B-A17B is a vision-language 397B MoE model with a native 262k-context window, multimodal input support, and strong reasoning.',
     },
+    {
+        id: 'grok-4.5',
+        aliases: ['grok-4.5-latest', 'grok-build-latest'],
+        provider: 'xai',
+        cost: {
+            input_per_million: 2.0,
+            output_per_million: 6.0,
+            cached_input_per_million: 0.5,
+        },
+        features: {
+            context_length: 500_000,
+            input_modality: ['text', 'image'],
+            output_modality: ['text'],
+            tool_use: true,
+            streaming: true,
+            json_output: true,
+            reasoning_output: true,
+        },
+        class: 'reasoning',
+        description: "Grok 4.5, xAI's flagship model for code and general chat with 500k context, multimodal input, tools, structured output, and configurable low/medium/high reasoning.",
+    },
     {
         id: 'grok-4.3',
         provider: 'xai',
@@ -518,7 +539,7 @@ exports.MODEL_REGISTRY = [
     },
     {
         id: 'grok-build-0.1',
-        aliases: ['grok-build', 'grok-build-latest'],
+        aliases: ['grok-build'],
         provider: 'xai',
         cost: {
             input_per_million: 1.0,
@@ -1802,7 +1823,7 @@ exports.MODEL_REGISTRY = [
     },
     {
         id: 'claude-sonnet-4-6',
-        aliases: ['claude-sonnet-4-6', 'claude-sonnet-4.6', 'claude-sonnet-latest', 'sonnet-4-6', 'sonnet-4.6'],
+        aliases: ['claude-sonnet-4-6', 'claude-sonnet-4.6', 'sonnet-4-6', 'sonnet-4.6'],
         provider: 'anthropic',
         cost: {
             input_per_million: 3.0,
@@ -1823,6 +1844,34 @@ exports.MODEL_REGISTRY = [
         score: 94,
         description: 'Claude Sonnet 4.6 with updated reasoning and multimodal capabilities.',
     },
+    {
+        id: 'claude-sonnet-5',
+        aliases: ['claude-sonnet-latest', 'sonnet-5', 'sonnet-5-latest'],
+        provider: 'anthropic',
+        cost: {
+            input_per_million: 2.0,
+            output_per_million: 10.0,
+            cached_input_per_million: 0.2,
+        },
+        features: {
+            context_length: 1_000_000,
+            input_modality: ['text', 'image'],
+            output_modality: ['text'],
+            tool_use: true,
+            streaming: true,
+            json_output: true,
+            reasoning_output: true,
+            max_output_tokens: 128000,
+        },
+        class: 'reasoning',
+        score: 96,
+        scores: {
+            monologue: 95,
+            code: 96,
+            reasoning: 95,
+        },
+        description: 'Claude Sonnet 5 - latest Sonnet model with 1M context, 128K output, vision support, and implicit adaptive thinking controls.',
+    },
     {
         id: 'claude-sonnet-4-5-20250514-long-context',
         aliases: ['claude-sonnet-4-5-long', 'claude-sonnet-4.5-long'],
@@ -2641,6 +2690,22 @@ exports.MODEL_REGISTRY = [
         class: 'voice',
         description: "Gemini's advanced text-to-speech model with superior voice quality, expression control, and multi-speaker support for creating dynamic conversations.",
     },
+    {
+        id: 'gemini-3.1-flash-tts-preview',
+        provider: 'google',
+        cost: {
+            input_per_million: 10.0,
+            output_per_million: 0,
+        },
+        features: {
+            input_modality: ['text'],
+            output_modality: ['audio'],
+            streaming: true,
+            context_length: 32000,
+        },
+        class: 'voice',
+        description: 'Gemini 3.1 Flash TTS preview for low-latency speech generation with steerable prompts and expressive audio tags.',
+    },
     {
         id: 'gemini-2.5-flash-native-audio-preview',
         aliases: ['gemini-2.5-flash-native-audio-preview-12-2025'],
@@ -2658,6 +2723,21 @@ exports.MODEL_REGISTRY = [
         class: 'voice',
         description: "Gemini's native audio preview model providing low-latency text-to-speech with built-in audio processing.",
     },
+    {
+        id: 'gemini-omni-flash',
+        provider: 'google',
+        cost: {
+            input_per_million: 0,
+            output_per_million: 0,
+        },
+        features: {
+            input_modality: ['text', 'image', 'audio', 'video'],
+            output_modality: ['text', 'image', 'audio'],
+            streaming: true,
+        },
+        class: 'standard',
+        description: 'Gemini Omni Flash preview for multimodal, conversational generation and editing across text, image, audio, and video inputs.',
+    },
     {
         id: 'perplexity/sonar',
         provider: 'openrouter',
@@ -3000,6 +3080,29 @@ exports.MODEL_REGISTRY = [
         class: 'transcription',
         description: 'Gemini Live API for real-time multimodal interaction with modality-specific pricing',
     },
+    {
+        id: 'gemini-3.1-flash-live-preview',
+        provider: 'google',
+        cost: {
+            input_per_million: {
+                text: 0.5,
+                audio: 3.0,
+                video: 3.0,
+            },
+            output_per_million: {
+                text: 2.0,
+                audio: 12.0,
+            },
+        },
+        features: {
+            context_length: 32000,
+            input_modality: ['text', 'audio', 'video'],
+            output_modality: ['text', 'audio'],
+            streaming: true,
+        },
+        class: 'transcription',
+        description: 'Gemini 3.1 Flash Live preview for high-quality, low-latency audio-to-audio realtime dialogue and voice-first applications.',
+    },
     {
         id: 'gemini-2.0-flash-live-001',
         provider: 'google',
@@ -3177,8 +3280,8 @@ exports.MODEL_REGISTRY = [
         provider: 'openrouter',
         openrouter_id: 'openai/gpt-oss-120b',
         cost: {
-            input_per_million: 0.039,
-            output_per_million: 0.18,
+            input_per_million: 0.03,
+            output_per_million: 0.15,
         },
         features: {
             context_length: 131072,
@@ -3237,6 +3340,29 @@ exports.MODEL_REGISTRY = [
         class: 'code',
         description: 'Qwen3.7 Max via OpenRouter. Flagship Qwen 3.7 model for agent-centric coding, productivity, and long-horizon autonomous execution.',
     },
+    {
+        id: 'qwen/qwen3.7-plus',
+        aliases: ['Qwen3.7 Plus', 'qwen3.7-plus', 'qwen-3.7-plus'],
+        provider: 'openrouter',
+        openrouter_id: 'qwen/qwen3.7-plus',
+        cost: {
+            input_per_million: 0.32,
+            cached_input_per_million: 0.064,
+            output_per_million: 1.28,
+        },
+        features: {
+            context_length: 1000000,
+            input_modality: ['text', 'image'],
+            output_modality: ['text'],
+            tool_use: true,
+            streaming: true,
+            json_output: true,
+            max_output_tokens: 65536,
+            reasoning_output: true,
+        },
+        class: 'reasoning',
+        description: 'Qwen3.7 Plus via OpenRouter. Current multimodal Qwen 3.7 model with 1M context for agentic workflows and visual reasoning.',
+    },
     {
         id: 'qwen/qwen3.6-plus',
         aliases: ['Qwen 3.6', 'Qwen3.6', 'qwen3.6', 'qwen-3.6', 'qwen3.6-plus', 'qwen-3.6-plus'],
@@ -3297,7 +3423,7 @@ exports.MODEL_REGISTRY = [
             tool_use: true,
             streaming: true,
             json_output: true,
-            max_output_tokens: 262140,
+            max_output_tokens: 262144,
             reasoning_output: true,
         },
         class: 'reasoning',
@@ -3332,8 +3458,9 @@ exports.MODEL_REGISTRY = [
         provider: 'openrouter',
         openrouter_id: 'qwen/qwen3.6-27b',
         cost: {
-            input_per_million: 0.29,
-            output_per_million: 3.2,
+            input_per_million: 0.285,
+            cached_input_per_million: 0.15,
+            output_per_million: 2.4,
         },
         features: {
             context_length: 262144,
@@ -3396,13 +3523,13 @@ exports.MODEL_REGISTRY = [
     },
     {
         id: 'z-ai/glm-5.1',
-        aliases: ['GLM-5.1', 'glm-5.1', 'glm-5', 'z-ai/glm-5'],
+        aliases: ['GLM-5.1', 'glm-5.1'],
         provider: 'openrouter',
         openrouter_id: 'z-ai/glm-5.1',
         cost: {
-            input_per_million: 0.98,
-            output_per_million: 3.08,
-            cached_input_per_million: 0.182,
+            input_per_million: 0.966,
+            output_per_million: 3.036,
+            cached_input_per_million: 0.1794,
         },
         features: {
             context_length: 202752,
@@ -3416,6 +3543,29 @@ exports.MODEL_REGISTRY = [
         class: 'reasoning',
         description: 'GLM-5.1 via OpenRouter. Z.ai flagship text model for long-horizon autonomous coding and sustained agent workflows.',
     },
+    {
+        id: 'z-ai/glm-5.2',
+        aliases: ['GLM-5.2', 'glm-5.2', 'glm-5'],
+        provider: 'openrouter',
+        openrouter_id: 'z-ai/glm-5.2',
+        cost: {
+            input_per_million: 0.93,
+            output_per_million: 3.0,
+            cached_input_per_million: 0.18,
+        },
+        features: {
+            context_length: 1048576,
+            input_modality: ['text'],
+            output_modality: ['text'],
+            tool_use: true,
+            streaming: true,
+            json_output: true,
+            max_output_tokens: 32768,
+            reasoning_output: true,
+        },
+        class: 'reasoning',
+        description: 'GLM-5.2 via OpenRouter. Z.ai 1M-context text model for long-horizon autonomous coding and sustained agent workflows.',
+    },
     {
         id: 'deepseek/deepseek-v4-pro',
         aliases: ['DeepSeek-V4', 'DeepSeek-V4-Pro', 'deepseek-v4', 'deepseek-v4-pro'],
@@ -3445,9 +3595,9 @@ exports.MODEL_REGISTRY = [
         provider: 'openrouter',
         openrouter_id: 'deepseek/deepseek-v4-flash',
         cost: {
-            input_per_million: 0.0983,
-            cached_input_per_million: 0.0197,
-            output_per_million: 0.1966,
+            input_per_million: 0.09,
+            cached_input_per_million: 0.018,
+            output_per_million: 0.18,
         },
         features: {
             context_length: 1048576,
@@ -3468,8 +3618,8 @@ exports.MODEL_REGISTRY = [
         provider: 'openrouter',
         openrouter_id: 'xiaomi/mimo-v2.5',
         cost: {
-            input_per_million: 0.14,
-            cached_input_per_million: 0.0028,
+            input_per_million: 0.105,
+            cached_input_per_million: 0.028,
             output_per_million: 0.28,
         },
         features: {
@@ -3533,14 +3683,14 @@ exports.MODEL_REGISTRY = [
         description: 'MiniMax M3 via OpenRouter. Multimodal 1M-context foundation model for long-horizon agentic work, coding, and video-aware reasoning.',
     },
     {
-        id: 'tencent/hy3-preview',
-        aliases: ['Tencent Hy3 Preview', 'Hy3 Preview', 'Hy3-preview', 'hy3-preview'],
+        id: 'tencent/hy3',
+        aliases: ['Tencent Hy3', 'Hy3', 'hy3', 'Tencent Hy3 Preview', 'Hy3 Preview', 'Hy3-preview', 'hy3-preview'],
         provider: 'openrouter',
-        openrouter_id: 'tencent/hy3-preview',
+        openrouter_id: 'tencent/hy3',
         cost: {
-            input_per_million: 0.063,
-            cached_input_per_million: 0.021,
-            output_per_million: 0.21,
+            input_per_million: 0.2,
+            cached_input_per_million: 0.5,
+            output_per_million: 0.8,
         },
         features: {
             context_length: 262144,
@@ -3549,19 +3699,20 @@ exports.MODEL_REGISTRY = [
             tool_use: true,
             streaming: true,
             json_output: false,
+            max_output_tokens: 131072,
             reasoning_output: true,
         },
         class: 'reasoning',
-        description: 'Tencent Hy3 preview via OpenRouter. High-efficiency MoE model for agentic workflows with configurable disabled, low, and high reasoning modes.',
+        description: 'Tencent Hy3 via OpenRouter. High-efficiency text model for agentic workflows with configurable disabled, low, and high reasoning modes.',
     },
     {
         id: 'moonshotai/kimi-k2.5',
         aliases: ['kimi-k2.5', 'kimi-k2-5'],
         provider: 'openrouter',
         cost: {
-            input_per_million: 0.4,
-            output_per_million: 1.9,
-            cached_input_per_million: 0.09,
+            input_per_million: 0.375,
+            output_per_million: 2.025,
+            cached_input_per_million: 0.203,
         },
         features: {
             context_length: 262144,
@@ -3582,9 +3733,9 @@ exports.MODEL_REGISTRY = [
         provider: 'openrouter',
         openrouter_id: 'moonshotai/kimi-k2.6',
         cost: {
-            input_per_million: 0.684,
-            output_per_million: 3.42,
-            cached_input_per_million: 0.144,
+            input_per_million: 0.66,
+            output_per_million: 3.41,
+            cached_input_per_million: 0.14,
         },
         features: {
             context_length: 262144,
@@ -3599,6 +3750,29 @@ exports.MODEL_REGISTRY = [
         class: 'reasoning',
         description: 'MoonshotAI Kimi K2.6 via OpenRouter. Multimodal long-horizon coding and multi-agent orchestration model with 262K context.',
     },
+    {
+        id: 'moonshotai/kimi-k2.7-code',
+        aliases: ['Kimi K2.7 Code', 'kimi-k2.7-code', 'kimi-k2-7-code', 'kimi-k2.7', 'kimi-k2-7'],
+        provider: 'openrouter',
+        openrouter_id: 'moonshotai/kimi-k2.7-code',
+        cost: {
+            input_per_million: 0.74,
+            output_per_million: 3.5,
+            cached_input_per_million: 0.15,
+        },
+        features: {
+            context_length: 262144,
+            input_modality: ['text', 'image'],
+            output_modality: ['text'],
+            tool_use: true,
+            streaming: true,
+            json_output: true,
+            max_output_tokens: 16384,
+            reasoning_output: true,
+        },
+        class: 'code',
+        description: 'MoonshotAI Kimi K2.7 Code via OpenRouter. Coding-focused long-context model for software engineering and agentic coding workflows.',
+    },
 ];
 function findModel(modelId) {
     const externalModel = (0, external_models_js_1.getExternalModel)(modelId);

package/dist/cjs/data/model_data.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"model_data.d.ts","sourceRoot":"","sources":["../../../data/model_data.ts"],"names":[],"mappings":"AAYA,OAAO,EACH,YAAY,EACZ,eAAe,EACf,WAAW,EACX,cAAc,EACd,aAAa,EACb,SAAS,EACT,aAAa,EACb,UAAU,EACV,UAAU,EACV,UAAU,EACb,MAAM,mBAAmB,CAAC;AAM3B,YAAY,EACR,YAAY,EACZ,eAAe,EACf,WAAW,EACX,cAAc,EACd,aAAa,EACb,SAAS,EACT,aAAa,EACb,UAAU,EACV,UAAU,EACV,UAAU,GACb,CAAC;AAOF,eAAO,MAAM,aAAa;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CA0MzB,CAAC;AAGF,eAAO,MAAM,cAAc,EAAE,UAAU,EA+lHtC,CAAC;AAQF,wBAAgB,SAAS,CAAC,OAAO,EAAE,MAAM,GAAG,UAAU,GAAG,SAAS,CAuBjE"}
+{"version":3,"file":"model_data.d.ts","sourceRoot":"","sources":["../../../data/model_data.ts"],"names":[],"mappings":"AAYA,OAAO,EACH,YAAY,EACZ,eAAe,EACf,WAAW,EACX,cAAc,EACd,aAAa,EACb,SAAS,EACT,aAAa,EACb,UAAU,EACV,UAAU,EACV,UAAU,EACb,MAAM,mBAAmB,CAAC;AAM3B,YAAY,EACR,YAAY,EACZ,eAAe,EACf,WAAW,EACX,cAAc,EACd,aAAa,EACb,SAAS,EACT,aAAa,EACb,UAAU,EACV,UAAU,EACV,UAAU,GACb,CAAC;AAOF,eAAO,MAAM,aAAa;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CA0MzB,CAAC;AAGF,eAAO,MAAM,cAAc,EAAE,UAAU,EA6xHtC,CAAC;AAQF,wBAAgB,SAAS,CAAC,OAAO,EAAE,MAAM,GAAG,UAAU,GAAG,SAAS,CAuBjE"}