From 0c8da28d02cc0cc529c476500c690e57c25d514f Mon Sep 17 00:00:00 2001 From: AstroHan Date: Thu, 24 Sep 2026 01:20:41 +0800 Subject: [PATCH 1/2] fix(model-metadata): follow the models.dev provider split and body-only override wave The weekly Model metadata upkeep workflow has failed three weeks in a row (2069716 / 3481615 / 3557136) before opening its review pull request: - models.dev split the kimi-for-coding provider into kimi-code-plan-cn and kimi-code-plan-global. The kimi-coding-plan access path now reads kimi-code-plan-cn, whose official endpoint is the kimi.com API URL the old provider carried, so configured connections keep their base URL. - fireworks-ai/models/minimax-m2p7 now ships a provider override with only a body member. Only npm selects a runtime adapter, so an override without npm projects to no row; unknown override keys still reject. Regenerating the snapshot surfaced the fallback-lifecycle contract: the deepseek and fireworks-ai fallback lists were serving models models.dev now marks deprecated, and the kimi-for-coding model row moved under the renamed provider. Fireworks derives its list, so it now filters deprecated ids like the zenmux derivation and anchors on the live kimi-k3; DeepSeek's hand list narrows to the current generation (deepseek-flash, deepseek-v4-pro). Offers change only; an id a user already chose still sends. The regenerated snapshot bundles the upstream wave this PR's registry edits reference (e.g. deepseek-flash only exists in the fresh catalog). Refs #4398 Generated-by: Maka (GLM-5.3-Flash) --- .../core/src/__tests__/model-metadata.test.ts | 6 +- packages/core/src/models-dev-projection.ts | 12 +- packages/core/src/provider-registry.ts | 12 +- .../models-dev-api.snapshot.json | 4543 ++++++++++++----- scripts/sync-model-metadata.mjs | 6 +- 5 files changed, 3301 insertions(+), 1278 deletions(-) diff --git a/packages/core/src/__tests__/model-metadata.test.ts b/packages/core/src/__tests__/model-metadata.test.ts index 2e8239947e..46a264caa4 100644 --- a/packages/core/src/__tests__/model-metadata.test.ts +++ b/packages/core/src/__tests__/model-metadata.test.ts @@ -205,10 +205,8 @@ describe('deepseek v4 flash vision exp metadata regression', () => { ); }); - it('keeps the model present in the deepseek shipped baseline', () => { - assert.ok( - providerFallbackModelIds(PROVIDER_REGISTRY.deepseek).includes('deepseek-v4-flash-vision-exp'), - ); + it('keeps the vision-capable baseline model in the deepseek shipped baseline', () => { + assert.ok(providerFallbackModelIds(PROVIDER_REGISTRY.deepseek).includes('deepseek-flash')); }); it('returns expected metadata from lookupModelMetadata', () => { diff --git a/packages/core/src/models-dev-projection.ts b/packages/core/src/models-dev-projection.ts index a2dc05e981..1d14405448 100644 --- a/packages/core/src/models-dev-projection.ts +++ b/packages/core/src/models-dev-projection.ts @@ -55,7 +55,7 @@ export const MODELS_DEV_PROVIDERS = { google: 'google', groq: 'groq', huggingface: 'huggingface', - 'kimi-coding-plan': 'kimi-for-coding', + 'kimi-coding-plan': 'kimi-code-plan-cn', MiniMax: 'minimax', 'MiniMax-cn': 'minimax-cn', 'minimax-coding-plan': 'minimax-coding-plan', @@ -109,7 +109,15 @@ export interface ModelsDevModel { readonly modalities?: { readonly input: string[]; readonly output: string[] }; readonly reasoning_options?: ReadonlyArray<{ readonly type?: string; readonly values?: unknown }>; readonly cost?: Readonly>; - readonly provider?: { readonly npm?: string; readonly api?: string }; + // Upstream models.dev entries are per-model hints keyed by provider id. An + // override without an `npm` field does not select a runtime adapter. The + // `body` member seen on some upstream entries carries request-body defaults + // that no downstream projection reads. + readonly provider?: { + readonly npm?: string; + readonly api?: string; + readonly body?: unknown; + }; } export type ModelsDevCatalog = Readonly>; diff --git a/packages/core/src/provider-registry.ts b/packages/core/src/provider-registry.ts index 7d9500717b..a38f08f63f 100644 --- a/packages/core/src/provider-registry.ts +++ b/packages/core/src/provider-registry.ts @@ -309,8 +309,8 @@ if (!fireworks.api) throw new Error('models.dev Fireworks AI provider facts are const fireworksModelIds = toolCallingModelIds( 'Fireworks AI', GENERATED_MODELS_DEV_METADATA['fireworks-ai'], - ['accounts/fireworks/models/kimi-k2p6'], -); + ['accounts/fireworks/models/kimi-k3'], +).filter((id) => GENERATED_MODELS_DEV_METADATA['fireworks-ai'][id]?.lifecycle !== 'deprecated'); const tencentTokenHub = GENERATED_MODELS_DEV_PROVIDER_FACTS['tencent-tokenhub']; if (tencentTokenHub.id !== 'tencent-tokenhub') { throw new Error( @@ -871,13 +871,7 @@ const providerRegistry = { label: 'DeepSeek', baseUrl: 'https://api.deepseek.com', authKind: 'api_key', - fallbackModels: [ - 'deepseek-v4-flash', - 'deepseek-v4-flash-vision-exp', - 'deepseek-v4-pro', - 'deepseek-reasoner', - 'deepseek-chat', - ], + fallbackModels: ['deepseek-flash', 'deepseek-v4-pro'], status: 'ready', runtimeAdapter: { kind: 'openai-compatible', diff --git a/scripts/model-metadata/models-dev-api.snapshot.json b/scripts/model-metadata/models-dev-api.snapshot.json index 6bc4b3f7c9..2485a42599 100644 --- a/scripts/model-metadata/models-dev-api.snapshot.json +++ b/scripts/model-metadata/models-dev-api.snapshot.json @@ -3,11 +3,11 @@ "sourceUrl": "https://models.dev/api.json", "origin": { "kind": "models-dev-response", - "retrievedAt": "2026-09-15T03:05:08.000Z", - "etag": "W/\"153362e16f62e2c0b4bbb4d66949e9c5\"", - "responseSha256": "153362e16f62e2c0b4bbb4d66949e9c5ccc050e0dd7dbd325bbf434f2a8fcbf1" + "retrievedAt": "2026-09-23T16:22:44.000Z", + "etag": "W/\"9ae55393e3711189a9fe8b32bd22d218\"", + "responseSha256": "9ae55393e3711189a9fe8b32bd22d218319de11c5c25fb27768ba8273f915eb7" }, - "projectionSha256": "da65200d9c31e4883f25bd3e4e6fe5fabe15b803472e85662761ba6950ef7ae8", + "projectionSha256": "c115e4fb78ec39b05ae3fcd25bd792cc31bf25a050e6c28279a817a6a699584a", "projection": { "metadata": { "anthropic": { @@ -224,6 +224,28 @@ "output": ["text"] } }, + "claude-opus-5-5": { + "displayName": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-06", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "claude-sonnet-4-5": { "displayName": "Claude Sonnet 4.5 (latest)", "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", @@ -353,6 +375,27 @@ "output": ["text"] } }, + "kimi-k3": { + "displayName": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 1048576, + "structuredOutput": true, + "lastUpdated": "2026-07-16", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "qvq-max": { "displayName": "QVQ Max", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", @@ -1612,6 +1655,29 @@ "output": ["text"] } }, + "deepseek-v4.1-flash": { + "displayName": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 384000, + "knowledgeCutoff": "2025-05", + "structuredOutput": true, + "lastUpdated": "2026-09-10", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"], + "toggle": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "glm-5": { "displayName": "GLM-5", "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", @@ -1671,6 +1737,27 @@ "output": ["text"] } }, + "glm-5.3": { + "displayName": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-08-14", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, "kimi-k2-thinking": { "displayName": "Moonshot Kimi K2 Thinking", "description": "Kimi reasoning model for long-horizon research, planning, and tool use", @@ -1733,6 +1820,27 @@ "output": ["text"] } }, + "kimi-k3": { + "displayName": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 1048576, + "structuredOutput": true, + "lastUpdated": "2026-07-16", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "kimi/kimi-k2.5": { "displayName": "kimi/kimi-k2.5", "description": "Kimi multimodal agent model for visual understanding, coding, and planning", @@ -2111,7 +2219,7 @@ "contextWindow": 34096, "maxOutputTokens": 4096, "knowledgeCutoff": "2024-04", - "lastUpdated": "2025-04-13", + "lastUpdated": "2026-09-11", "capabilities": { "vision": true, "reasoning": false, @@ -2501,7 +2609,7 @@ "contextWindow": 1048576, "maxOutputTokens": 65536, "knowledgeCutoff": "2025-04", - "lastUpdated": "2025-07-23", + "lastUpdated": "2026-09-11", "capabilities": { "vision": false, "reasoning": false, @@ -2519,12 +2627,16 @@ "contextWindow": 262144, "maxOutputTokens": 65536, "knowledgeCutoff": "2025-04", - "lastUpdated": "2025-09-23", + "structuredOutput": true, + "lastUpdated": "2026-09-22", "capabilities": { "vision": false, - "reasoning": false, + "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "toggle": true + }, "modalities": { "input": ["text"], "output": ["text"] @@ -2691,7 +2803,7 @@ "maxOutputTokens": 65536, "knowledgeCutoff": "2025-04", "structuredOutput": true, - "lastUpdated": "2026-02-23", + "lastUpdated": "2026-09-22", "capabilities": { "vision": true, "reasoning": true, @@ -2712,7 +2824,8 @@ "contextWindow": 1000000, "maxOutputTokens": 65536, "knowledgeCutoff": "2025-04", - "lastUpdated": "2026-02-16", + "structuredOutput": true, + "lastUpdated": "2026-09-22", "capabilities": { "vision": true, "reasoning": true, @@ -3708,6 +3821,27 @@ "output": ["text"] } }, + "glm-5.3": { + "displayName": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-08-14", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, "happyhorse-1.1-i2v": { "displayName": "HappyHorse 1.1 Image-to-Video", "description": "Video model for image-to-video generation", @@ -4167,6 +4301,29 @@ "output": ["text"] } }, + "deepseek-v4.1-flash": { + "displayName": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 384000, + "knowledgeCutoff": "2025-05", + "structuredOutput": false, + "lastUpdated": "2026-09-10", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"], + "toggle": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "glm-5": { "displayName": "GLM-5", "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", @@ -4230,6 +4387,27 @@ "output": ["text"] } }, + "glm-5.3": { + "displayName": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-08-14", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, "happyhorse-1.1-i2v": { "displayName": "HappyHorse 1.1 Image-to-Video", "description": "Video model for image-to-video generation", @@ -4883,7 +5061,7 @@ "cloudflare-workers-ai": { "@cf/aisingapore/gemma-sea-lion-v4-27b-it": { "displayName": "Gemma Sea Lion V4 27B It", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "description": "Gemma 3 27B tuned by AI Singapore for Southeast Asian languages and instruction following", "lifecycle": "active", "contextWindow": 128000, "maxOutputTokens": 128000, @@ -4933,8 +5111,7 @@ "functionCalling": true }, "thinkingOptions": { - "efforts": ["high", "max"], - "toggle": true + "efforts": ["none", "low", "high", "max"] }, "modalities": { "input": ["text"], @@ -4955,8 +5132,7 @@ "functionCalling": true }, "thinkingOptions": { - "efforts": ["high", "max"], - "toggle": true + "efforts": ["none", "low", "high", "max"] }, "modalities": { "input": ["text"], @@ -4977,8 +5153,7 @@ "functionCalling": true }, "thinkingOptions": { - "efforts": ["low", "medium", "high"], - "toggle": true + "efforts": ["none", "high"] }, "modalities": { "input": ["text", "image"], @@ -4992,7 +5167,7 @@ "contextWindow": 131000, "maxOutputTokens": 131000, "structuredOutput": false, - "lastUpdated": "2025-10-07", + "lastUpdated": "2025-10-02", "capabilities": { "vision": false, "reasoning": false, @@ -5142,8 +5317,9 @@ "lifecycle": "active", "contextWindow": 128000, "maxOutputTokens": 128000, + "knowledgeCutoff": "2024-06", "structuredOutput": false, - "lastUpdated": "2025-03-18", + "lastUpdated": "2025-03-17", "capabilities": { "vision": false, "reasoning": false, @@ -5169,8 +5345,7 @@ "functionCalling": true }, "thinkingOptions": { - "efforts": ["low", "medium", "high"], - "toggle": true + "efforts": ["none", "high"] }, "modalities": { "input": ["text", "image"], @@ -5191,10 +5366,6 @@ "reasoning": true, "functionCalling": true }, - "thinkingOptions": { - "efforts": ["low", "medium", "high"], - "toggle": true - }, "modalities": { "input": ["text", "image"], "output": ["text"] @@ -5256,6 +5427,9 @@ "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "efforts": ["low", "medium", "high"] + }, "modalities": { "input": ["text"], "output": ["text"] @@ -5353,7 +5527,6 @@ "functionCalling": true }, "thinkingOptions": { - "efforts": ["low", "medium", "high"], "toggle": true }, "modalities": { @@ -5375,8 +5548,7 @@ "functionCalling": true }, "thinkingOptions": { - "efforts": ["low", "medium", "high"], - "toggle": true + "efforts": ["none", "high", "max"] }, "modalities": { "input": ["text"], @@ -5388,7 +5560,7 @@ "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "lifecycle": "active", "contextWindow": 1310720, - "maxOutputTokens": 1310720, + "maxOutputTokens": 1048576, "structuredOutput": true, "lastUpdated": "2026-08-14", "capabilities": { @@ -5397,7 +5569,7 @@ "functionCalling": true }, "thinkingOptions": { - "efforts": ["low", "medium", "high"] + "efforts": ["low", "high", "max"] }, "modalities": { "input": ["text"], @@ -5417,6 +5589,9 @@ "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, "modalities": { "input": ["text", "image"], "output": ["text"] @@ -5631,7 +5806,7 @@ "contextWindow": 1048576, "maxOutputTokens": 384000, "structuredOutput": true, - "lastUpdated": "2026-08-21", + "lastUpdated": "2026-09-01", "capabilities": { "vision": true, "reasoning": true, @@ -6795,7 +6970,7 @@ "deepseek-v4-flash": { "displayName": "DeepSeek V4 Flash", "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "lifecycle": "active", + "lifecycle": "deprecated", "contextWindow": 1000000, "maxOutputTokens": 384000, "knowledgeCutoff": "2025-05", @@ -6818,7 +6993,7 @@ "deepseek-v4-flash-vision-exp": { "displayName": "DeepSeek V4 Flash Vision Exp", "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "lifecycle": "active", + "lifecycle": "deprecated", "contextWindow": 1000000, "maxOutputTokens": 384000, "knowledgeCutoff": "2025-05", @@ -6852,7 +7027,7 @@ "functionCalling": true }, "thinkingOptions": { - "efforts": ["high", "max"], + "efforts": ["low", "high", "max"], "toggle": true }, "modalities": { @@ -6865,7 +7040,7 @@ "accounts/fireworks/models/deepseek-v4-flash-0731": { "displayName": "DeepSeek V4 Flash 0731", "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "lifecycle": "active", + "lifecycle": "deprecated", "contextWindow": 1000000, "maxOutputTokens": 384000, "knowledgeCutoff": "2025-05", @@ -6888,11 +7063,11 @@ "accounts/fireworks/models/deepseek-v4-flash-vision-exp": { "displayName": "DeepSeek V4 Flash Vision Exp", "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "lifecycle": "active", + "lifecycle": "deprecated", "contextWindow": 1000000, "maxOutputTokens": 384000, "structuredOutput": true, - "lastUpdated": "2026-08-21", + "lastUpdated": "2026-09-01", "capabilities": { "vision": true, "reasoning": true, @@ -6907,10 +7082,33 @@ "output": ["text"] } }, + "accounts/fireworks/models/deepseek-v4-pro": { + "displayName": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "lifecycle": "deprecated", + "contextWindow": 1000000, + "maxOutputTokens": 384000, + "knowledgeCutoff": "2025-05", + "structuredOutput": true, + "lastUpdated": "2026-04-24", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["high", "max"], + "toggle": true + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, "accounts/fireworks/models/deepseek-v4-pro-0813": { "displayName": "DeepSeek V4 Pro 0813", "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "lifecycle": "active", + "lifecycle": "deprecated", "contextWindow": 1000000, "maxOutputTokens": 384000, "structuredOutput": true, @@ -6955,7 +7153,7 @@ "accounts/fireworks/models/glm-5p2": { "displayName": "GLM 5.2", "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "lifecycle": "active", + "lifecycle": "deprecated", "contextWindow": 1048575, "maxOutputTokens": 131072, "lastUpdated": "2026-06-16", @@ -7011,7 +7209,7 @@ "efforts": ["low", "high", "max"] }, "modalities": { - "input": ["text", "image", "video", "pdf"], + "input": ["text", "image"], "output": ["text"] } }, @@ -7055,7 +7253,7 @@ "accounts/fireworks/models/kimi-k2p6": { "displayName": "Kimi K2.6", "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "lifecycle": "active", + "lifecycle": "deprecated", "contextWindow": 262000, "maxOutputTokens": 262000, "lastUpdated": "2026-04-17", @@ -7075,7 +7273,7 @@ "accounts/fireworks/models/kimi-k2p7-code": { "displayName": "Kimi K2.7 Code", "description": "Kimi coding model for software agents, refactors, and repository reasoning", - "lifecycle": "active", + "lifecycle": "deprecated", "contextWindow": 262000, "maxOutputTokens": 262000, "lastUpdated": "2026-06-16", @@ -7114,15 +7312,15 @@ "output": ["text"] } }, - "accounts/fireworks/models/minimax-m3": { - "displayName": "MiniMax-M3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "lifecycle": "active", - "contextWindow": 512000, - "maxOutputTokens": 512000, - "lastUpdated": "2026-06-12", + "accounts/fireworks/models/minimax-m2p7": { + "displayName": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "lifecycle": "deprecated", + "contextWindow": 196608, + "maxOutputTokens": 131072, + "lastUpdated": "2026-03-18", "capabilities": { - "vision": true, + "vision": false, "reasoning": true, "functionCalling": true }, @@ -7130,32 +7328,34 @@ "efforts": ["low", "medium", "high"] }, "modalities": { - "input": ["text", "image", "video"], + "input": ["text"], "output": ["text"] } }, - "accounts/fireworks/models/mistral-large-3-fp8": { - "displayName": "Mistral Large 3 675B Instruct 2512", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "accounts/fireworks/models/minimax-m3": { + "displayName": "MiniMax-M3", + "description": "Fireworks text-only MiniMax coding model for long-context reasoning and agent tasks", "lifecycle": "active", - "contextWindow": 262144, - "maxOutputTokens": 262144, - "knowledgeCutoff": "2024-11", - "lastUpdated": "2025-12-02", + "contextWindow": 512000, + "maxOutputTokens": 512000, + "lastUpdated": "2026-06-12", "capabilities": { - "vision": true, - "reasoning": false, + "vision": false, + "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "efforts": ["low", "medium", "high"] + }, "modalities": { - "input": ["text", "image"], + "input": ["text"], "output": ["text"] } }, "accounts/fireworks/models/muse-glimmer-30b": { "displayName": "Muse Glimmer 30B", "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "lifecycle": "active", + "lifecycle": "deprecated", "contextWindow": 131072, "maxOutputTokens": 131072, "knowledgeCutoff": "2026-01-04", @@ -7264,12 +7464,57 @@ "contextWindow": 262144, "maxOutputTokens": 131072, "lastUpdated": "2026-08-03", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, + "accounts/fireworks/routers/deepseek-flash-latest": { + "displayName": "DeepSeek Flash Latest", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 384000, + "knowledgeCutoff": "2025-05", + "structuredOutput": true, + "lastUpdated": "2026-09-10", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"], + "toggle": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, + "accounts/fireworks/routers/deepseek-pro-latest": { + "displayName": "DeepSeek Pro Latest", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 384000, + "structuredOutput": true, + "lastUpdated": "2026-08-22", "capabilities": { "vision": false, "reasoning": true, "functionCalling": true }, "thinkingOptions": { + "efforts": ["high", "max"], "toggle": true }, "modalities": { @@ -7319,6 +7564,91 @@ "output": ["text"] } }, + "accounts/fireworks/routers/glm-fast-latest": { + "displayName": "GLM 5.3 Fast (Latest)", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "lifecycle": "active", + "contextWindow": 1048572, + "maxOutputTokens": 262144, + "structuredOutput": true, + "lastUpdated": "2026-09-15", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, + "accounts/fireworks/routers/glm-flash-latest": { + "displayName": "GLM Flash Latest (GLM 5.3 Flash)", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "lifecycle": "active", + "contextWindow": 1048573, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-09-15", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, + "accounts/fireworks/routers/glm-latest": { + "displayName": "GLM Latest", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "lifecycle": "active", + "contextWindow": 1048573, + "maxOutputTokens": 262144, + "structuredOutput": true, + "lastUpdated": "2026-09-15", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, + "accounts/fireworks/routers/kimi-fast-latest": { + "displayName": "Kimi Fast Latest", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-09-15", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "max"], + "toggle": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "accounts/fireworks/routers/kimi-k3-fast": { "displayName": "Kimi K3 Fast", "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", @@ -7340,6 +7670,68 @@ "input": ["text", "image"], "output": ["text"] } + }, + "accounts/fireworks/routers/kimi-latest": { + "displayName": "Kimi Latest", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-09-15", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "max"], + "toggle": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, + "accounts/fireworks/routers/minimax-latest": { + "displayName": "MiniMax Latest", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "lifecycle": "active", + "contextWindow": 512000, + "maxOutputTokens": 512000, + "lastUpdated": "2026-09-15", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high"] + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, + "accounts/fireworks/routers/qwen-max-latest": { + "displayName": "Qwen Max Latest (Qwen3.8 Max)", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "lifecycle": "active", + "contextWindow": 262144, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-16", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } } }, "github-copilot": { @@ -7471,6 +7863,27 @@ "output": ["text"] } }, + "claude-opus-5.5": { + "displayName": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-06", + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "claude-sonnet-4.6": { "displayName": "Claude Sonnet 4.6", "description": "Claude workhorse for coding agents, careful analysis, and production cost control", @@ -7832,6 +8245,52 @@ "output": ["text"] } }, + "gpt-6-luna": { + "displayName": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "lifecycle": "active", + "contextWindow": 1050000, + "inputLimit": 922000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-05-18", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, + "gpt-6-sol": { + "displayName": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "lifecycle": "active", + "contextWindow": 1050000, + "inputLimit": 922000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-04-20", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "grok-4.5": { "displayName": "Grok 4.5", "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", @@ -7877,6 +8336,29 @@ "output": ["text"] } }, + "grok-4.7": { + "displayName": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "lifecycle": "active", + "contextWindow": 500000, + "inputLimit": 372000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-05", + "structuredOutput": true, + "lastUpdated": "2026-09-21", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "kimi-k2.7-code": { "displayName": "Kimi K2.7 Code", "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", @@ -9193,7 +9675,7 @@ "contextWindow": 1048576, "maxOutputTokens": 384000, "structuredOutput": true, - "lastUpdated": "2026-08-21", + "lastUpdated": "2026-09-01", "capabilities": { "vision": true, "reasoning": true, @@ -10173,6 +10655,27 @@ "output": ["text"] } }, + "tencent/Hy4-preview": { + "displayName": "Hy4 preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 64000, + "structuredOutput": true, + "lastUpdated": "2026-08-28", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "high"] + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, "thinkingmachines/Inkling": { "displayName": "Inkling", "description": "Multimodal model for analyzing text, images, documents, and rich media", @@ -11568,7 +12071,28 @@ "functionCalling": true }, "thinkingOptions": { - "efforts": ["high", "max"] + "efforts": ["none", "high", "max"] + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, + "zai-glm-5-3": { + "displayName": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-08-14", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] }, "modalities": { "input": ["text"], @@ -13635,6 +14159,27 @@ "output": ["text"] } }, + "z-ai/glm-5.3": { + "displayName": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-08-14", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, "z-ai/glm-5.3-flash": { "displayName": "GLM-5.3-Flash", "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", @@ -14880,6 +15425,52 @@ "output": ["text"] } }, + "gpt-6-luna": { + "displayName": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "lifecycle": "active", + "contextWindow": 1050000, + "inputLimit": 922000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-05-18", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, + "gpt-6-sol": { + "displayName": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "lifecycle": "active", + "contextWindow": 1050000, + "inputLimit": 922000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-04-20", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "gpt-image-1": { "displayName": "gpt-image-1", "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", @@ -15372,6 +15963,27 @@ "output": ["text"] } }, + "claude-opus-5-5": { + "displayName": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-06", + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "claude-sonnet-4": { "displayName": "Claude Sonnet 4", "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", @@ -15502,7 +16114,7 @@ "contextWindow": 1000000, "maxOutputTokens": 384000, "structuredOutput": true, - "lastUpdated": "2026-08-21", + "lastUpdated": "2026-09-01", "capabilities": { "vision": true, "reasoning": true, @@ -15540,6 +16152,28 @@ "output": ["text"] } }, + "deepseek-v4.1-flash": { + "displayName": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 384000, + "knowledgeCutoff": "2025-05", + "structuredOutput": true, + "lastUpdated": "2026-09-10", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "gemini-3-flash": { "displayName": "Gemini 3 Flash", "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", @@ -16319,7 +16953,7 @@ } }, "gpt-5.6-sol": { - "displayName": "GPT-5.6 Sol (50% Off)", + "displayName": "GPT-5.6 Sol", "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", "lifecycle": "active", "contextWindow": 1050000, @@ -16387,6 +17021,52 @@ "output": ["text"] } }, + "gpt-6-luna": { + "displayName": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "lifecycle": "active", + "contextWindow": 1050000, + "inputLimit": 922000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-05-18", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, + "gpt-6-sol": { + "displayName": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "lifecycle": "active", + "contextWindow": 1050000, + "inputLimit": 922000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-04-20", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "grok-4.5": { "displayName": "Grok 4.5", "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", @@ -16430,6 +17110,28 @@ "output": ["text"] } }, + "grok-4.7": { + "displayName": "Grok 4.7 (30% Off)", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "lifecycle": "active", + "contextWindow": 500000, + "maxOutputTokens": 500000, + "knowledgeCutoff": "2026-05", + "structuredOutput": true, + "lastUpdated": "2026-09-21", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "grok-build-0.1": { "displayName": "Grok Build 0.1", "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", @@ -16821,7 +17523,7 @@ "mimo-v2.5-free": { "displayName": "MiMo V2.5 Free", "description": "MiMo omni model for text, image, video, audio, and agents", - "lifecycle": "active", + "lifecycle": "deprecated", "contextWindow": 200000, "maxOutputTokens": 32000, "knowledgeCutoff": "2024-12", @@ -16836,6 +17538,23 @@ "output": ["text"] } }, + "mimo-v2.6-flash-free": { + "displayName": "MiMo-V2.6-Flash Free", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "lifecycle": "active", + "contextWindow": 200000, + "maxOutputTokens": 32000, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } + }, "minimax-m2.1": { "displayName": "MiniMax-M2.1", "description": "Legacy model retained for compatibility with older integrations", @@ -17020,7 +17739,7 @@ "functionCalling": true }, "thinkingOptions": { - "efforts": ["minimal", "low", "medium", "high", "xhigh", "max"] + "efforts": ["minimal", "low", "medium", "high", "xhigh"] }, "modalities": { "input": ["text", "image", "video", "pdf", "audio"], @@ -17205,6 +17924,28 @@ "output": ["text"] } }, + "qwen3.8-flash": { + "displayName": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-08-26", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "xhigh"], + "toggle": true + }, + "modalities": { + "input": ["text", "image", "video"], + "output": ["text"] + } + }, "ring-2.6-1t-free": { "displayName": "Ring 2.6 1T Free", "description": "Legacy model retained for compatibility with older integrations", @@ -17223,6 +17964,27 @@ "output": ["text"] } }, + "space-bunny-free": { + "displayName": "Space Bunny Free", + "description": "Anonymous preview reasoning model for coding, agentic tasks, tool use, and multimodal input", + "lifecycle": "active", + "contextWindow": 1048576, + "inputLimit": 524288, + "maxOutputTokens": 524288, + "lastUpdated": "2026-09-23", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "video"], + "output": ["text"] + } + }, "trinity-large-preview-free": { "displayName": "Trinity Large Preview", "description": "Legacy model retained for compatibility with older integrations", @@ -17293,7 +18055,7 @@ "contextWindow": 1000000, "maxOutputTokens": 384000, "structuredOutput": true, - "lastUpdated": "2026-08-21", + "lastUpdated": "2026-09-01", "capabilities": { "vision": true, "reasoning": true, @@ -17517,6 +18279,28 @@ "output": ["text"] } }, + "grok-4.7": { + "displayName": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "lifecycle": "active", + "contextWindow": 500000, + "maxOutputTokens": 500000, + "knowledgeCutoff": "2026-05", + "structuredOutput": true, + "lastUpdated": "2026-09-21", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "hy3": { "displayName": "Hy3", "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", @@ -17726,6 +18510,40 @@ "output": ["text"] } }, + "mimo-v2.6-flash": { + "displayName": "MiMo-V2.6-Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } + }, + "mimo-v2.6-pro": { + "displayName": "MiMo-V2.6-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } + }, "minimax-m2.5": { "displayName": "MiniMax-M2.5", "description": "Legacy model retained for compatibility with older integrations", @@ -17992,6 +18810,27 @@ "input": ["text", "image", "video"], "output": ["text"] } + }, + "space-bunny-free": { + "displayName": "Space Bunny Free", + "description": "Anonymous preview reasoning model for coding, agentic tasks, tool use, and multimodal input", + "lifecycle": "active", + "contextWindow": 1048576, + "inputLimit": 524288, + "maxOutputTokens": 524288, + "lastUpdated": "2026-09-23", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "video"], + "output": ["text"] + } } }, "openrouter": { @@ -18051,8 +18890,7 @@ "functionCalling": true }, "thinkingOptions": { - "efforts": ["low", "medium", "high", "xhigh", "max"], - "toggle": true + "efforts": ["low", "medium", "high", "xhigh", "max"] }, "modalities": { "input": ["text", "image", "pdf"], @@ -18109,7 +18947,7 @@ "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", "lifecycle": "active", "contextWindow": 1048576, - "maxOutputTokens": 943718, + "maxOutputTokens": 393216, "structuredOutput": true, "lastUpdated": "2026-09-14", "capabilities": { @@ -18131,7 +18969,7 @@ "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", "lifecycle": "active", "contextWindow": 1310720, - "maxOutputTokens": 393216, + "maxOutputTokens": 943718, "structuredOutput": true, "lastUpdated": "2026-08-01", "capabilities": { @@ -18440,6 +19278,48 @@ "output": ["text"] } }, + "aion-labs/aion-3.5": { + "displayName": "Aion 3.5", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "lifecycle": "active", + "contextWindow": 262144, + "maxOutputTokens": 32768, + "structuredOutput": false, + "lastUpdated": "2026-09-23", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, + "aion-labs/aion-3.5-mini": { + "displayName": "Aion 3.5 Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "lifecycle": "active", + "contextWindow": 262144, + "maxOutputTokens": 32768, + "structuredOutput": false, + "lastUpdated": "2026-09-23", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, "aion-labs/aion-rp-llama-3.1-8b": { "displayName": "Aion-RP 1.0 (8B)", "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", @@ -18659,28 +19539,6 @@ "output": ["text"] } }, - "anthropic/claude-opus-4": { - "displayName": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "lifecycle": "active", - "contextWindow": 200000, - "maxOutputTokens": 32000, - "knowledgeCutoff": "2025-01-31", - "structuredOutput": false, - "lastUpdated": "2025-05-22", - "capabilities": { - "vision": true, - "reasoning": true, - "functionCalling": true - }, - "thinkingOptions": { - "toggle": true - }, - "modalities": { - "input": ["image", "text", "pdf"], - "output": ["text"] - } - }, "anthropic/claude-opus-4.1": { "displayName": "Claude Opus 4.1 (latest)", "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", @@ -18817,11 +19675,33 @@ "output": ["text"] } }, + "anthropic/claude-opus-5.5": { + "displayName": "Claude Opus 5.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-06", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "anthropic/claude-sonnet-4": { "displayName": "Claude Sonnet 4", "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "lifecycle": "active", - "contextWindow": 1000000, + "contextWindow": 200000, "maxOutputTokens": 64000, "knowledgeCutoff": "2025-01-31", "structuredOutput": false, @@ -19133,6 +20013,27 @@ "output": ["text"] } }, + "cohere/command-a-plus": { + "displayName": "Command A+", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "lifecycle": "active", + "contextWindow": 192000, + "maxOutputTokens": 64000, + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "cohere/command-r-08-2024": { "displayName": "Command R", "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", @@ -19216,7 +20117,7 @@ "description": "DeepSeek chat model for instruction following, coding, and analysis", "lifecycle": "active", "contextWindow": 163840, - "maxOutputTokens": 16000, + "maxOutputTokens": 16384, "knowledgeCutoff": "2025-09", "structuredOutput": true, "lastUpdated": "2026-02-28", @@ -19450,7 +20351,7 @@ "contextWindow": 1048576, "maxOutputTokens": 943718, "structuredOutput": true, - "lastUpdated": "2026-08-21", + "lastUpdated": "2026-09-01", "capabilities": { "vision": true, "reasoning": true, @@ -19470,7 +20371,7 @@ "description": "Open MoE flagship with million-token context for coding and long agent runs", "lifecycle": "active", "contextWindow": 1048576, - "maxOutputTokens": 393216, + "maxOutputTokens": 384000, "knowledgeCutoff": "2025-05", "structuredOutput": true, "lastUpdated": "2026-04-24", @@ -19515,7 +20416,7 @@ "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", "lifecycle": "active", "contextWindow": 1048576, - "maxOutputTokens": 384000, + "maxOutputTokens": 943718, "knowledgeCutoff": "2025-05", "structuredOutput": true, "lastUpdated": "2026-09-10", @@ -20442,24 +21343,6 @@ "output": ["text"] } }, - "kwaipilot/kat-coder-pro-v2": { - "displayName": "KAT-Coder-Pro V2", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "lifecycle": "active", - "contextWindow": 262144, - "maxOutputTokens": 144000, - "structuredOutput": true, - "lastUpdated": "2026-03-27", - "capabilities": { - "vision": false, - "reasoning": false, - "functionCalling": true - }, - "modalities": { - "input": ["text"], - "output": ["text"] - } - }, "kwaipilot/kat-coder-pro-v2.5": { "displayName": "KAT-Coder-Pro V2.5", "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", @@ -20693,7 +21576,7 @@ "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", "lifecycle": "active", "contextWindow": 131072, - "maxOutputTokens": 117964, + "maxOutputTokens": 16384, "knowledgeCutoff": "2026-01-04", "structuredOutput": true, "lastUpdated": "2026-08-10", @@ -21135,25 +22018,6 @@ "output": ["text"] } }, - "mistralai/mistral-large-2512": { - "displayName": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "lifecycle": "active", - "contextWindow": 262144, - "maxOutputTokens": 209715, - "knowledgeCutoff": "2024-11", - "structuredOutput": true, - "lastUpdated": "2025-12-02", - "capabilities": { - "vision": true, - "reasoning": false, - "functionCalling": true - }, - "modalities": { - "input": ["text", "image", "pdf"], - "output": ["text"] - } - }, "mistralai/mistral-medium-3": { "displayName": "Mistral Medium 3", "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", @@ -21304,7 +22168,7 @@ "capabilities": { "vision": true, "reasoning": false, - "functionCalling": false + "functionCalling": true }, "modalities": { "input": ["text", "image"], @@ -21545,6 +22409,27 @@ "output": ["text"] } }, + "nex-agi/nex-n2.5-mini": { + "displayName": "Nex-N2.5-Mini", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "lifecycle": "active", + "contextWindow": 262144, + "maxOutputTokens": 235929, + "structuredOutput": true, + "lastUpdated": "2026-09-08", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": false + }, + "thinkingOptions": { + "efforts": ["none", "medium", "high"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "nex-agi/nex-n2.5-mini:free": { "displayName": "Nex-N2.5-Mini (free)", "description": "Efficient model for low-latency assistance, extraction, and routine automation", @@ -21566,6 +22451,27 @@ "output": ["text"] } }, + "nex-agi/nex-n2.5-pro": { + "displayName": "Nex-N2.5-Pro", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "lifecycle": "active", + "contextWindow": 262144, + "maxOutputTokens": 235929, + "structuredOutput": true, + "lastUpdated": "2026-09-08", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "medium", "high"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "nex-agi/nex-n2.5-pro:free": { "displayName": "Nex-N2.5-Pro (free)", "description": "Multimodal reasoning model for visual analysis, planning, and tool use", @@ -22820,6 +23726,94 @@ "output": ["text"] } }, + "openai/gpt-6-luna": { + "displayName": "GPT-6 Luna", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "lifecycle": "active", + "contextWindow": 1050000, + "inputLimit": 922000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-05-18", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, + "openai/gpt-6-luna-pro": { + "displayName": "GPT-6 Luna Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "lifecycle": "active", + "contextWindow": 1050000, + "maxOutputTokens": 128000, + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["pdf", "image", "text"], + "output": ["text"] + } + }, + "openai/gpt-6-sol": { + "displayName": "GPT-6 Sol", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "lifecycle": "active", + "contextWindow": 1050000, + "inputLimit": 922000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-04-20", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, + "openai/gpt-6-sol-pro": { + "displayName": "GPT-6 Sol Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "lifecycle": "active", + "contextWindow": 1050000, + "maxOutputTokens": 128000, + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["pdf", "image", "text"], + "output": ["text"] + } + }, "openai/gpt-audio": { "displayName": "GPT Audio", "description": "Speech generation model for controllable voice, narration, and audio delivery", @@ -22879,7 +23873,7 @@ "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", "lifecycle": "active", "contextWindow": 131072, - "maxOutputTokens": 117964, + "maxOutputTokens": 65536, "structuredOutput": true, "lastUpdated": "2025-08-05", "capabilities": { @@ -22900,7 +23894,7 @@ "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", "lifecycle": "active", "contextWindow": 131072, - "maxOutputTokens": 117964, + "maxOutputTokens": 32768, "structuredOutput": true, "lastUpdated": "2025-08-05", "capabilities": { @@ -23405,6 +24399,28 @@ "output": ["text"] } }, + "prism-ml/ternary-bonsai-2-27b": { + "displayName": "Ternary Bonsai 2 27B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "lifecycle": "active", + "contextWindow": 262144, + "maxOutputTokens": 32768, + "structuredOutput": true, + "lastUpdated": "2026-09-18", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["medium", "xhigh"], + "toggle": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "qwen/qwen-2.5-72b-instruct": { "displayName": "Qwen2.5 72B Instruct", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", @@ -23863,7 +24879,7 @@ "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", "lifecycle": "active", "contextWindow": 262144, - "maxOutputTokens": 32768, + "maxOutputTokens": 235929, "knowledgeCutoff": "2025-04", "structuredOutput": true, "lastUpdated": "2025-09", @@ -23920,7 +24936,7 @@ "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "lifecycle": "active", "contextWindow": 262144, - "maxOutputTokens": 16384, + "maxOutputTokens": 32768, "knowledgeCutoff": "2025-03-31", "structuredOutput": true, "lastUpdated": "2025-10-06", @@ -24054,7 +25070,7 @@ "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "lifecycle": "active", "contextWindow": 262144, - "maxOutputTokens": 65536, + "maxOutputTokens": 16384, "structuredOutput": true, "lastUpdated": "2026-02-23", "capabilities": { @@ -24096,7 +25112,7 @@ "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "lifecycle": "active", "contextWindow": 262144, - "maxOutputTokens": 235929, + "maxOutputTokens": 32768, "structuredOutput": true, "lastUpdated": "2026-02-23", "capabilities": { @@ -24181,7 +25197,7 @@ "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "lifecycle": "active", "contextWindow": 262144, - "maxOutputTokens": 65536, + "maxOutputTokens": 262140, "structuredOutput": true, "lastUpdated": "2026-04-22", "capabilities": { @@ -24391,6 +25407,28 @@ "output": ["text"] } }, + "qwen/qwen3.8-27b:free": { + "displayName": "Qwen3.8 27B (free)", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "lifecycle": "active", + "contextWindow": 262144, + "maxOutputTokens": 235929, + "structuredOutput": true, + "lastUpdated": "2026-08-14", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "xhigh"], + "toggle": true + }, + "modalities": { + "input": ["text", "image", "video"], + "output": ["text"] + } + }, "qwen/qwen3.8-flash": { "displayName": "Qwen3.8 Flash", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", @@ -24433,6 +25471,27 @@ "output": ["text"] } }, + "qwen/qwen3.8-omni-flash": { + "displayName": "Qwen3.8 Omni Flash", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-09-17", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } + }, "rekaai/reka-edge": { "displayName": "Reka Edge", "description": "Multimodal model for analyzing text, images, documents, and rich media", @@ -24880,7 +25939,7 @@ "capabilities": { "vision": false, "reasoning": false, - "functionCalling": true + "functionCalling": false }, "modalities": { "input": ["text"], @@ -24971,6 +26030,24 @@ "output": ["text"] } }, + "unbiased/pareto": { + "displayName": "Pareto", + "description": "Blended multimodal model that runs multiple language models in parallel and returns a synthesized answer", + "lifecycle": "active", + "contextWindow": 262144, + "maxOutputTokens": 131072, + "structuredOutput": false, + "lastUpdated": "2026-09-17", + "capabilities": { + "vision": true, + "reasoning": false, + "functionCalling": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "undi95/remm-slerp-l2-13b": { "displayName": "ReMM SLERP 13B", "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", @@ -24990,6 +26067,27 @@ "output": ["text"] } }, + "upstage/solar-mini4": { + "displayName": "Solar Mini 4", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "lifecycle": "active", + "contextWindow": 524288, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-09-23", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "minimal", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, "upstage/solar-pro-3": { "displayName": "Solar Pro 3", "description": "Flagship model for demanding analysis, coding, and production agent workflows", @@ -25004,7 +26102,7 @@ "functionCalling": true }, "thinkingOptions": { - "toggle": true + "efforts": ["none", "minimal", "low", "medium", "high"] }, "modalities": { "input": ["text"], @@ -25025,7 +26123,7 @@ "functionCalling": true }, "thinkingOptions": { - "toggle": true + "efforts": ["none", "minimal", "low", "medium", "high", "xhigh", "max"] }, "modalities": { "input": ["text"], @@ -25158,6 +26256,28 @@ "output": ["text"] } }, + "x-ai/grok-4.7": { + "displayName": "Grok 4.7", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "lifecycle": "active", + "contextWindow": 500000, + "maxOutputTokens": 450000, + "knowledgeCutoff": "2026-05", + "structuredOutput": true, + "lastUpdated": "2026-09-21", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "x-ai/grok-build-0.1": { "displayName": "Grok Build 0.1", "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", @@ -25220,6 +26340,69 @@ "output": ["text"] } }, + "xiaomi/mimo-v2.6-flash": { + "displayName": "MiMo-V2.6-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } + }, + "xiaomi/mimo-v2.6-pro": { + "displayName": "MiMo-V2.6-Pro", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } + }, + "xiaomi/mimo-v2.6-pro-ultraspeed": { + "displayName": "MiMo-V2.6-Pro-UltraSpeed", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-09-21", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } + }, "z-ai/glm-4.5": { "displayName": "GLM-4.5", "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", @@ -25442,7 +26625,7 @@ "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "lifecycle": "active", "contextWindow": 1048576, - "maxOutputTokens": 128000, + "maxOutputTokens": 131072, "structuredOutput": true, "lastUpdated": "2026-06-13", "capabilities": { @@ -25459,12 +26642,34 @@ "output": ["text"] } }, + "z-ai/glm-5.2:free": { + "displayName": "GLM 5.2 (free)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "lifecycle": "active", + "contextWindow": 32768, + "maxOutputTokens": 29491, + "structuredOutput": false, + "lastUpdated": "2026-06-13", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": false + }, + "thinkingOptions": { + "efforts": ["high", "xhigh"], + "toggle": true + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, "z-ai/glm-5.3": { "displayName": "GLM-5.3", "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "lifecycle": "active", "contextWindow": 1310720, - "maxOutputTokens": 943717, + "maxOutputTokens": 131072, "structuredOutput": true, "lastUpdated": "2026-08-14", "capabilities": { @@ -25485,7 +26690,7 @@ "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "lifecycle": "active", "contextWindow": 1310720, - "maxOutputTokens": 131072, + "maxOutputTokens": 943718, "structuredOutput": true, "lastUpdated": "2026-08-26", "capabilities": { @@ -25501,6 +26706,27 @@ "output": ["text"] } }, + "z-ai/glm-5.3-flashx": { + "displayName": "GLM 5.3 FlashX", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "structuredOutput": false, + "lastUpdated": "2026-09-18", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text", "image", "video"], + "output": ["text"] + } + }, "z-ai/glm-5v-turbo": { "displayName": "GLM-5V-Turbo", "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", @@ -25524,24 +26750,6 @@ } }, "siliconflow": { - "baidu/ERNIE-4.5-300B-A47B": { - "displayName": "baidu/ERNIE-4.5-300B-A47B", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "lifecycle": "active", - "contextWindow": 131000, - "maxOutputTokens": 131000, - "structuredOutput": true, - "lastUpdated": "2025-11-25", - "capabilities": { - "vision": false, - "reasoning": false, - "functionCalling": true - }, - "modalities": { - "input": ["text"], - "output": ["text"] - } - }, "ByteDance-Seed/Seed-OSS-36B-Instruct": { "displayName": "ByteDance-Seed/Seed-OSS-36B-Instruct", "description": "Tool-capable chat model for instruction following and agentic application workflows", @@ -25699,6 +26907,43 @@ "output": ["text"] } }, + "deepseek-ai/DeepSeek-V4-Flash-0731": { + "displayName": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 384000, + "knowledgeCutoff": "2025-05", + "structuredOutput": true, + "lastUpdated": "2026-07-31", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, + "deepseek-ai/DeepSeek-V4-Flash-Vision-Exp": { + "displayName": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 384000, + "structuredOutput": true, + "lastUpdated": "2026-09-01", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "deepseek-ai/DeepSeek-V4-Pro": { "displayName": "DeepSeek V4 Pro", "description": "Open MoE flagship with million-token context for coding and long agent runs", @@ -25718,6 +26963,42 @@ "output": ["text"] } }, + "deepseek-ai/DeepSeek-V4-Pro-0813": { + "displayName": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 384000, + "structuredOutput": true, + "lastUpdated": "2026-08-22", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, + "google/gemma-4-12B-it": { + "displayName": "Gemma 4 12B IT", + "description": "Compact Gemma 4 instruction model for open, self-hosted chat and reasoning", + "lifecycle": "active", + "contextWindow": 262144, + "maxOutputTokens": 262144, + "structuredOutput": true, + "lastUpdated": "2026-06-09", + "capabilities": { + "vision": false, + "reasoning": false, + "functionCalling": true + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, "google/gemma-4-26B-A4B-it": { "displayName": "Gemma 4 26B A4B IT", "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", @@ -25772,6 +27053,23 @@ "output": ["text"] } }, + "meituan-longcat/LongCat-2.0": { + "displayName": "LongCat-2.0", + "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", + "lifecycle": "active", + "contextWindow": 1049000, + "maxOutputTokens": 131072, + "lastUpdated": "2026-06-30", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, "MiniMaxAI/MiniMax-M2.5": { "displayName": "MiniMaxAI/MiniMax-M2.5", "description": "MiniMax model for chat, coding, office work, and agentic tasks", @@ -25790,6 +27088,23 @@ "output": ["text"] } }, + "MiniMaxAI/MiniMax-M3": { + "displayName": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131000, + "lastUpdated": "2026-06-01", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "moonshotai/Kimi-K2.5": { "displayName": "moonshotai/Kimi-K2.5", "description": "Kimi multimodal agent model for visual understanding, coding, and planning", @@ -25826,6 +27141,61 @@ "output": ["text"] } }, + "moonshotai/Kimi-K2.7-Code": { + "displayName": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "lifecycle": "active", + "contextWindow": 262144, + "maxOutputTokens": 262144, + "knowledgeCutoff": "2025-01", + "structuredOutput": true, + "lastUpdated": "2026-06-12", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, + "moonshotai/Kimi-K3": { + "displayName": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 262000, + "structuredOutput": true, + "lastUpdated": "2026-07-16", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, + "nex-agi/Nex-N2-Pro": { + "displayName": "Nex-N2-Pro", + "description": "Open agentic MoE model (397B total, 17B active) for coding, tool use, and research workflows", + "lifecycle": "active", + "contextWindow": 262144, + "maxOutputTokens": 256000, + "structuredOutput": true, + "lastUpdated": "2026-06-02", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "openai/gpt-oss-120b": { "displayName": "openai/gpt-oss-120b", "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", @@ -25919,24 +27289,6 @@ "output": ["text"] } }, - "Qwen/Qwen3-235B-A22B-Thinking-2507": { - "displayName": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "lifecycle": "active", - "contextWindow": 262000, - "maxOutputTokens": 262000, - "structuredOutput": true, - "lastUpdated": "2025-11-25", - "capabilities": { - "vision": false, - "reasoning": true, - "functionCalling": true - }, - "modalities": { - "input": ["text"], - "output": ["text"] - } - }, "Qwen/Qwen3-30B-A3B-Instruct-2507": { "displayName": "Qwen/Qwen3-30B-A3B-Instruct-2507", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", @@ -26033,42 +27385,6 @@ "output": ["text"] } }, - "Qwen/Qwen3-VL-235B-A22B-Instruct": { - "displayName": "Qwen/Qwen3-VL-235B-A22B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "lifecycle": "active", - "contextWindow": 262000, - "maxOutputTokens": 262000, - "structuredOutput": true, - "lastUpdated": "2025-11-25", - "capabilities": { - "vision": true, - "reasoning": false, - "functionCalling": true - }, - "modalities": { - "input": ["text", "image"], - "output": ["text"] - } - }, - "Qwen/Qwen3-VL-235B-A22B-Thinking": { - "displayName": "Qwen/Qwen3-VL-235B-A22B-Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "lifecycle": "active", - "contextWindow": 262000, - "maxOutputTokens": 262000, - "structuredOutput": true, - "lastUpdated": "2025-11-25", - "capabilities": { - "vision": true, - "reasoning": true, - "functionCalling": true - }, - "modalities": { - "input": ["text", "image"], - "output": ["text"] - } - }, "Qwen/Qwen3-VL-30B-A3B-Instruct": { "displayName": "Qwen/Qwen3-VL-30B-A3B-Instruct", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", @@ -26285,6 +27601,24 @@ "output": ["text"] } }, + "Qwen/Qwen3.8-2.4T-A95B": { + "displayName": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "lifecycle": "active", + "contextWindow": 1049000, + "maxOutputTokens": 131000, + "structuredOutput": true, + "lastUpdated": "2026-08-12", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, "stepfun-ai/Step-3.5-Flash": { "displayName": "stepfun-ai/Step-3.5-Flash", "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", @@ -26324,13 +27658,14 @@ "output": ["text"] } }, - "tencent/Hy3-preview": { - "displayName": "Hy3 preview", + "tencent/Hy3": { + "displayName": "Hy3", "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", "lifecycle": "active", "contextWindow": 262144, + "inputLimit": 192000, "maxOutputTokens": 262144, - "lastUpdated": "2026-04-20", + "lastUpdated": "2026-07-06", "capabilities": { "vision": false, "reasoning": true, @@ -26382,7 +27717,7 @@ "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "lifecycle": "active", "contextWindow": 205000, - "maxOutputTokens": 205000, + "maxOutputTokens": 131000, "structuredOutput": true, "lastUpdated": "2026-04-08", "capabilities": { @@ -26416,6 +27751,48 @@ "output": ["text"] } }, + "zai-org/GLM-5.3": { + "displayName": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "lifecycle": "active", + "contextWindow": 1049000, + "maxOutputTokens": 262000, + "structuredOutput": true, + "lastUpdated": "2026-08-14", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, + "zai-org/GLM-5.3-Flash": { + "displayName": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "lifecycle": "active", + "contextWindow": 1049000, + "maxOutputTokens": 262000, + "structuredOutput": true, + "lastUpdated": "2026-08-26", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "zai-org/GLM-5V-Turbo": { "displayName": "zai-org/GLM-5V-Turbo", "description": "GLM vision model for visual reasoning, documents, and multimodal agents", @@ -26542,6 +27919,28 @@ "output": ["text"] } }, + "step-5-preview": { + "displayName": "Step 5 Preview", + "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", + "lifecycle": "active", + "contextWindow": 1000000, + "inputLimit": 1000000, + "maxOutputTokens": 1000000, + "structuredOutput": true, + "lastUpdated": "2026-09-20", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high"] + }, + "modalities": { + "input": ["text", "image", "video"], + "output": ["text"] + } + }, "step-tts-2": { "displayName": "Step TTS 2", "description": "Speech generation model for controllable voice, narration, and audio delivery", @@ -26693,6 +28092,28 @@ "output": ["text"] } }, + "step-5-preview": { + "displayName": "Step 5 Preview", + "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", + "lifecycle": "active", + "contextWindow": 1000000, + "inputLimit": 1000000, + "maxOutputTokens": 1000000, + "structuredOutput": true, + "lastUpdated": "2026-09-20", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high"] + }, + "modalities": { + "input": ["text", "image", "video"], + "output": ["text"] + } + }, "step-tts-2": { "displayName": "Step TTS 2", "description": "Speech generation model for controllable voice, narration, and audio delivery", @@ -26805,6 +28226,28 @@ "input": ["text", "image", "video"], "output": ["text"] } + }, + "step-5-preview": { + "displayName": "Step 5 Preview", + "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", + "lifecycle": "active", + "contextWindow": 1000000, + "inputLimit": 1000000, + "maxOutputTokens": 1000000, + "structuredOutput": true, + "lastUpdated": "2026-09-20", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high"] + }, + "modalities": { + "input": ["text", "image", "video"], + "output": ["text"] + } } }, "stepfun-step-plan": { @@ -26874,6 +28317,28 @@ "output": ["text"] } }, + "step-5-preview": { + "displayName": "Step 5 Preview", + "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", + "lifecycle": "active", + "contextWindow": 1000000, + "inputLimit": 1000000, + "maxOutputTokens": 1000000, + "structuredOutput": true, + "lastUpdated": "2026-09-20", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high"] + }, + "modalities": { + "input": ["text", "image", "video"], + "output": ["text"] + } + }, "step-router-v1": { "displayName": "Step Router v1", "description": "StepFun routing model that dispatches requests to the appropriate Step model.", @@ -28220,8 +29685,8 @@ "displayName": "Qwen3 Next 80B A3B Instruct", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "lifecycle": "active", - "contextWindow": 131072, - "maxOutputTokens": 32768, + "contextWindow": 262114, + "maxOutputTokens": 262114, "knowledgeCutoff": "2025-04", "lastUpdated": "2025-09", "capabilities": { @@ -28238,8 +29703,8 @@ "displayName": "Qwen3 Next 80B A3B Thinking", "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", "lifecycle": "active", - "contextWindow": 131072, - "maxOutputTokens": 32768, + "contextWindow": 262144, + "maxOutputTokens": 262144, "knowledgeCutoff": "2025-04", "lastUpdated": "2025-09", "capabilities": { @@ -28517,7 +29982,7 @@ "displayName": "Qwen 3.8 Max", "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", "lifecycle": "active", - "contextWindow": 1000000, + "contextWindow": 262144, "maxOutputTokens": 128000, "lastUpdated": "2026-07-19", "capabilities": { @@ -28554,6 +30019,26 @@ "output": ["text"] } }, + "alibaba/qwen3.8-omni-flash": { + "displayName": "Qwen 3.8 Omni Flash", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-17", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "minimal", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "alibaba/wan-v2.5-t2v-preview": { "displayName": "Wan v2.5 Text-to-Video Preview", "description": "Video model for prompt-guided generation, editing, and motion workflows", @@ -28884,7 +30369,7 @@ "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "lifecycle": "active", "contextWindow": 200000, - "maxOutputTokens": 8192, + "maxOutputTokens": 32000, "knowledgeCutoff": "2025-03-31", "lastUpdated": "2025-05-22", "capabilities": { @@ -29048,12 +30533,54 @@ "output": ["text"] } }, + "anthropic/claude-opus-5.5": { + "displayName": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-06", + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, + "anthropic/claude-opus-5.5-fast": { + "displayName": "Claude Opus 5.5 (Fast)", + "description": "Claude model for long-running agentic coding and knowledge work", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-06", + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "anthropic/claude-sonnet-4": { "displayName": "Claude Sonnet 4", "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "lifecycle": "active", "contextWindow": 1000000, - "maxOutputTokens": 8192, + "maxOutputTokens": 64000, "knowledgeCutoff": "2025-03-31", "lastUpdated": "2025-05-22", "capabilities": { @@ -29779,7 +31306,7 @@ "contextWindow": 1048576, "maxOutputTokens": 1048576, "structuredOutput": true, - "lastUpdated": "2026-08-21", + "lastUpdated": "2026-09-01", "capabilities": { "vision": true, "reasoning": true, @@ -29843,8 +31370,8 @@ "displayName": "DeepSeek V4.1 Flash", "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", "lifecycle": "active", - "contextWindow": 1000000, - "maxOutputTokens": 384000, + "contextWindow": 1048576, + "maxOutputTokens": 32768, "knowledgeCutoff": "2025-05", "structuredOutput": true, "lastUpdated": "2026-09-10", @@ -29877,21 +31404,6 @@ "output": ["audio"] } }, - "fish-audio/s1-free": { - "displayName": "S1 (Free)", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "lifecycle": "active", - "lastUpdated": "2025-10-20", - "capabilities": { - "vision": false, - "reasoning": false, - "functionCalling": false - }, - "modalities": { - "input": ["text"], - "output": ["audio"] - } - }, "fish-audio/s2-pro": { "displayName": "S2 Pro", "description": "Speech generation model for controllable voice, narration, and audio delivery", @@ -29907,21 +31419,6 @@ "output": ["audio"] } }, - "fish-audio/s2-pro-free": { - "displayName": "S2 Pro (Free)", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "lifecycle": "active", - "lastUpdated": "2026-03-09", - "capabilities": { - "vision": false, - "reasoning": false, - "functionCalling": false - }, - "modalities": { - "input": ["text"], - "output": ["audio"] - } - }, "fish-audio/s2.1-pro": { "displayName": "S2.1 Pro", "description": "Speech generation model for controllable voice, narration, and audio delivery", @@ -29937,21 +31434,6 @@ "output": ["audio"] } }, - "fish-audio/s2.1-pro-free": { - "displayName": "S2.1 Pro (Free)", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "lifecycle": "active", - "lastUpdated": "2026-07-28", - "capabilities": { - "vision": false, - "reasoning": false, - "functionCalling": false - }, - "modalities": { - "input": ["text"], - "output": ["audio"] - } - }, "fish-audio/transcribe-1": { "displayName": "Transcribe-1", "description": "Speech transcription model for accurate audio-to-text and captioning workflows", @@ -29967,21 +31449,6 @@ "output": ["text"] } }, - "fish-audio/transcribe-1-free": { - "displayName": "Transcribe-1 (Free)", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "lifecycle": "active", - "lastUpdated": "2026-03-01", - "capabilities": { - "vision": false, - "reasoning": false, - "functionCalling": false - }, - "modalities": { - "input": ["audio"], - "output": ["text"] - } - }, "google/gemini-2.5-flash": { "displayName": "Gemini 2.5 Flash", "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", @@ -30009,8 +31476,8 @@ "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", "lifecycle": "active", "contextWindow": 32768, - "maxOutputTokens": 65536, - "knowledgeCutoff": "2025-01", + "maxOutputTokens": 65535, + "knowledgeCutoff": "2024-06", "lastUpdated": "2025-08-26", "capabilities": { "vision": true, @@ -30027,7 +31494,7 @@ "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", "lifecycle": "active", "contextWindow": 1048576, - "maxOutputTokens": 65536, + "maxOutputTokens": 65535, "knowledgeCutoff": "2025-01", "structuredOutput": true, "lastUpdated": "2025-06-17", @@ -30308,7 +31775,7 @@ "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "lifecycle": "active", "contextWindow": 1000000, - "maxOutputTokens": 65536, + "maxOutputTokens": 65535, "knowledgeCutoff": "2026-03", "structuredOutput": true, "lastUpdated": "2026-08-13", @@ -30330,7 +31797,7 @@ "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", "lifecycle": "active", "contextWindow": 1000000, - "maxOutputTokens": 65536, + "maxOutputTokens": 65535, "structuredOutput": true, "lastUpdated": "2026-09-02", "capabilities": { @@ -30343,6 +31810,36 @@ "output": ["text"] } }, + "google/gemini-3.8-live": { + "displayName": "Gemini 3.8 Live", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "lifecycle": "active", + "lastUpdated": "2026-09-15", + "capabilities": { + "vision": false, + "reasoning": false, + "functionCalling": false + }, + "modalities": { + "input": ["text", "audio"], + "output": ["text", "audio"] + } + }, + "google/gemini-3.8-live-extended-thinking": { + "displayName": "Gemini 3.8 Live Extended Thinking", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "lifecycle": "active", + "lastUpdated": "2026-09-15", + "capabilities": { + "vision": false, + "reasoning": false, + "functionCalling": false + }, + "modalities": { + "input": ["text", "audio"], + "output": ["text", "audio"] + } + }, "google/gemini-embedding-001": { "displayName": "Gemini Embedding 001", "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", @@ -30727,6 +32224,40 @@ "output": ["text"] } }, + "inference-net/schematron-v2-small": { + "displayName": "Schematron V2 Small", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "lifecycle": "active", + "contextWindow": 128000, + "maxOutputTokens": 4096, + "lastUpdated": "2026-04-16", + "capabilities": { + "vision": false, + "reasoning": false, + "functionCalling": false + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, + "inference-net/schematron-v2-turbo": { + "displayName": "Schematron V2 Turbo", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "lifecycle": "active", + "contextWindow": 128000, + "maxOutputTokens": 8192, + "lastUpdated": "2026-04-16", + "capabilities": { + "vision": false, + "reasoning": false, + "functionCalling": false + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, "interfaze/interfaze-beta": { "displayName": "Interfaze Beta", "description": "Multimodal reasoning model for visual analysis, planning, and tool use", @@ -30867,75 +32398,6 @@ "output": ["video"] } }, - "kwaipilot/kat-coder-air-v2.5": { - "displayName": "Kat Coder Air V2.5", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "lifecycle": "active", - "contextWindow": 256000, - "maxOutputTokens": 80000, - "lastUpdated": "2026-07-10", - "capabilities": { - "vision": true, - "reasoning": true, - "functionCalling": true - }, - "modalities": { - "input": ["text", "image"], - "output": ["text"] - } - }, - "kwaipilot/kat-coder-pro-v1": { - "displayName": "KAT-Coder-Pro V1", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "lifecycle": "active", - "contextWindow": 256000, - "maxOutputTokens": 32000, - "knowledgeCutoff": "2024-10", - "lastUpdated": "2025-10-24", - "capabilities": { - "vision": false, - "reasoning": true, - "functionCalling": false - }, - "modalities": { - "input": ["text"], - "output": ["text"] - } - }, - "kwaipilot/kat-coder-pro-v2": { - "displayName": "Kat Coder Pro V2", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "lifecycle": "active", - "contextWindow": 256000, - "maxOutputTokens": 256000, - "lastUpdated": "2026-03-30", - "capabilities": { - "vision": false, - "reasoning": true, - "functionCalling": true - }, - "modalities": { - "input": ["text"], - "output": ["text"] - } - }, - "kwaipilot/kat-coder-pro-v2.5": { - "displayName": "Kat Coder Pro V2.5", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "lifecycle": "active", - "contextWindow": 256000, - "maxOutputTokens": 80000, - "lastUpdated": "2026-07-10", - "capabilities": { - "vision": true, - "reasoning": true, - "functionCalling": true - }, - "modalities": { - "input": ["text", "image"], - "output": ["text"] - } - }, "meta/llama-3.1-70b": { "displayName": "Llama 3.1 70B Instruct", "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", @@ -31369,47 +32831,11 @@ "output": ["text"] } }, - "mistral/devstral-2": { - "displayName": "Devstral 2", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "lifecycle": "active", - "contextWindow": 256000, - "maxOutputTokens": 256000, - "knowledgeCutoff": "2024-10", - "lastUpdated": "2025-12-09", - "capabilities": { - "vision": false, - "reasoning": false, - "functionCalling": true - }, - "modalities": { - "input": ["text"], - "output": ["text"] - } - }, - "mistral/devstral-small-2": { - "displayName": "Devstral Small 2", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "lifecycle": "active", - "contextWindow": 256000, - "maxOutputTokens": 256000, - "knowledgeCutoff": "2024-10", - "lastUpdated": "2025-05-07", - "capabilities": { - "vision": true, - "reasoning": false, - "functionCalling": true - }, - "modalities": { - "input": ["text", "image"], - "output": ["text"] - } - }, "mistral/ministral-14b": { "displayName": "Ministral 14B", "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", "lifecycle": "active", - "contextWindow": 256000, + "contextWindow": 262144, "maxOutputTokens": 256000, "knowledgeCutoff": "2024-10", "lastUpdated": "2025-12-01", @@ -31480,7 +32906,7 @@ "displayName": "Mistral Large 3", "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", "lifecycle": "active", - "contextWindow": 256000, + "contextWindow": 262144, "maxOutputTokens": 256000, "knowledgeCutoff": "2024-10", "lastUpdated": "2025-12-02", @@ -31494,29 +32920,11 @@ "output": ["text"] } }, - "mistral/mistral-medium": { - "displayName": "Mistral Medium 3.1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "lifecycle": "active", - "contextWindow": 128000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2024-10", - "lastUpdated": "2025-05-07", - "capabilities": { - "vision": true, - "reasoning": false, - "functionCalling": true - }, - "modalities": { - "input": ["text", "image"], - "output": ["text"] - } - }, "mistral/mistral-medium-3.5": { "displayName": "Mistral Medium Latest", "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", "lifecycle": "active", - "contextWindow": 256000, + "contextWindow": 262144, "maxOutputTokens": 256000, "lastUpdated": "2026-05-21", "capabilities": { @@ -31536,17 +32944,17 @@ "displayName": "Mistral Nemo", "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", "lifecycle": "active", - "contextWindow": 128000, - "maxOutputTokens": 128000, + "contextWindow": 60288, + "maxOutputTokens": 16000, "knowledgeCutoff": "2024-04", "lastUpdated": "2024-07-01", "capabilities": { - "vision": true, + "vision": false, "reasoning": false, "functionCalling": true }, "modalities": { - "input": ["text", "image"], + "input": ["text"], "output": ["text"] } }, @@ -31554,7 +32962,7 @@ "displayName": "Mistral Small (latest)", "description": "Efficient Mistral model for fast chat, extraction, and production assistants", "lifecycle": "active", - "contextWindow": 32000, + "contextWindow": 262144, "maxOutputTokens": 4000, "knowledgeCutoff": "2025-06", "lastUpdated": "2026-03-16", @@ -31568,21 +32976,20 @@ "output": ["text"] } }, - "mistral/pixtral-12b": { - "displayName": "Pixtral 12B", - "description": "Mistral vision-language model for image understanding and multimodal chat", + "mixedbread/toast-1": { + "displayName": "Toast 1", + "description": "Specialized search model for knowledge-intensive questions, multi-step retrieval, and evidence synthesis", "lifecycle": "active", - "contextWindow": 128000, - "maxOutputTokens": 128000, - "knowledgeCutoff": "2024-09", - "lastUpdated": "2024-09-01", + "contextWindow": 131000, + "maxOutputTokens": 4000, + "lastUpdated": "2026-08-13", "capabilities": { - "vision": true, + "vision": false, "reasoning": false, "functionCalling": true }, "modalities": { - "input": ["text", "image"], + "input": ["text"], "output": ["text"] } }, @@ -31625,8 +33032,8 @@ "displayName": "Kimi K2.5", "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", "lifecycle": "active", - "contextWindow": 262114, - "maxOutputTokens": 262114, + "contextWindow": 256000, + "maxOutputTokens": 256000, "knowledgeCutoff": "2025-01", "structuredOutput": true, "lastUpdated": "2026-01", @@ -32959,6 +34366,98 @@ "output": ["text"] } }, + "openai/gpt-6-luna": { + "displayName": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "lifecycle": "active", + "contextWindow": 1050000, + "inputLimit": 922000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-05-18", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, + "openai/gpt-6-luna-fast": { + "displayName": "GPT-6 Luna (Fast)", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "lifecycle": "active", + "contextWindow": 1050000, + "inputLimit": 922000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-05-18", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, + "openai/gpt-6-sol": { + "displayName": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "lifecycle": "active", + "contextWindow": 1050000, + "inputLimit": 922000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-04-20", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, + "openai/gpt-6-sol-fast": { + "displayName": "GPT-6 Sol (Fast)", + "description": "OpenAI model for complex coding and agentic workflows", + "lifecycle": "active", + "contextWindow": 1050000, + "inputLimit": 922000, + "maxOutputTokens": 128000, + "knowledgeCutoff": "2026-04-20", + "structuredOutput": true, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "openai/gpt-image-1": { "displayName": "GPT Image 1", "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", @@ -33049,13 +34548,27 @@ "output": ["image"] } }, + "openai/gpt-live-1": { + "displayName": "GPT-Live 1", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "lifecycle": "active", + "lastUpdated": "2026-09-10", + "capabilities": { + "vision": false, + "reasoning": false, + "functionCalling": false + }, + "modalities": { + "input": ["text", "audio"], + "output": ["text", "audio"] + } + }, "openai/gpt-oss-120b": { "displayName": "GPT OSS 120B", "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", "lifecycle": "active", "contextWindow": 131072, "maxOutputTokens": 131072, - "knowledgeCutoff": "2024-10", "structuredOutput": true, "lastUpdated": "2025-08-05", "capabilities": { @@ -33078,7 +34591,6 @@ "contextWindow": 131072, "inputLimit": 122880, "maxOutputTokens": 8192, - "knowledgeCutoff": "2024-10", "structuredOutput": true, "lastUpdated": "2025-08-05", "capabilities": { @@ -33641,6 +35153,46 @@ "output": ["image"] } }, + "quiverai/arrow-2": { + "displayName": "Arrow 2", + "description": "Fast SVG generation model for creation, vectorization, editing, and animation", + "lifecycle": "active", + "contextWindow": 131072, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-16", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text", "image"] + } + }, + "quiverai/arrow-2-telos": { + "displayName": "Arrow 2 Telos", + "description": "High-fidelity SVG generation model for complex vector work and long-context refinement", + "lifecycle": "active", + "contextWindow": 131072, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-16", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text", "image"] + } + }, "recraft/recraft-v2": { "displayName": "Recraft V2", "description": "Image model for prompt-driven generation, editing, and visual design workflows", @@ -34039,6 +35591,28 @@ "output": ["text"] } }, + "spacexai/grok-4.7": { + "displayName": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "lifecycle": "active", + "contextWindow": 500000, + "maxOutputTokens": 500000, + "knowledgeCutoff": "2026-05", + "structuredOutput": true, + "lastUpdated": "2026-09-21", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "spacexai/grok-build-0.1": { "displayName": "Grok Build 0.1", "description": "Grok coding model for agentic engineering, edits, and codebase workflows", @@ -34220,6 +35794,25 @@ "output": ["text"] } }, + "stepfun/step-5-preview": { + "displayName": "Step 5 Preview", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "lifecycle": "active", + "contextWindow": 1000000, + "inputLimit": 1000000, + "maxOutputTokens": 1000000, + "structuredOutput": true, + "lastUpdated": "2026-09-20", + "capabilities": { + "vision": true, + "reasoning": false, + "functionCalling": false + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "tencent/hy-mt2-lite": { "displayName": "Tencent Hy-MT2-Lite", "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", @@ -34594,6 +36187,66 @@ "output": ["text"] } }, + "xiaomi/mimo-v2.6-flash": { + "displayName": "MiMo V2.6 Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "minimal", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, + "xiaomi/mimo-v2.6-pro": { + "displayName": "MiMo V2.6 Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "minimal", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, + "xiaomi/mimo-v2.6-pro-ultraspeed": { + "displayName": "MiMo V2.6 Pro UltraSpeed", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-21", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["none", "minimal", "low", "medium", "high", "xhigh", "max"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "zai/glm-4.5": { "displayName": "GLM 4.5", "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", @@ -34910,6 +36563,27 @@ "output": ["text"] } }, + "zai/glm-5.3-flashx": { + "displayName": "GLM 5.3 FlashX", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-08-26", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "zai/glm-5v-turbo": { "displayName": "GLM 5V Turbo", "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", @@ -35053,6 +36727,28 @@ "output": ["text"] } }, + "grok-4.7": { + "displayName": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "lifecycle": "active", + "contextWindow": 500000, + "maxOutputTokens": 500000, + "knowledgeCutoff": "2026-05", + "structuredOutput": true, + "lastUpdated": "2026-09-21", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh"] + }, + "modalities": { + "input": ["text", "image", "pdf"], + "output": ["text"] + } + }, "grok-build-0.1": { "displayName": "Grok Build 0.1", "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", @@ -35087,25 +36783,9 @@ "output": ["image", "pdf"] } }, - "grok-imagine-image-2.0": { - "displayName": "Grok Imagine Image 2.0", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "lifecycle": "active", - "contextWindow": 64000, - "lastUpdated": "2026-08-07", - "capabilities": { - "vision": true, - "reasoning": false, - "functionCalling": false - }, - "modalities": { - "input": ["text", "image", "pdf"], - "output": ["image", "pdf"] - } - }, "grok-imagine-image-quality": { "displayName": "Grok Imagine Image Quality", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "description": "Higher-fidelity Grok Imagine image model for prompt-driven generation, editing, and visual design workflows", "lifecycle": "active", "contextWindow": 16000, "lastUpdated": "2026-04-03", @@ -35278,6 +36958,66 @@ "input": ["text"], "output": ["text"] } + }, + "mimo-v2.6-flash": { + "displayName": "MiMo-V2.6-Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } + }, + "mimo-v2.6-pro": { + "displayName": "MiMo-V2.6-Pro", + "description": "MiMo Pro model for multimodal coding agents and long-context automation", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } + }, + "mimo-v2.6-pro-ultraspeed": { + "displayName": "MiMo-V2.6-Pro-UltraSpeed", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-21", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } } }, "xiaomi-token-plan-cn": { @@ -35411,55 +37151,187 @@ "input": ["text"], "output": ["audio"] } - } - }, - "xiaomi-token-plan-sgp": { - "mimo-v2-pro": { - "displayName": "MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "lifecycle": "deprecated", - "contextWindow": 1048576, - "maxOutputTokens": 131072, - "knowledgeCutoff": "2024-12", - "lastUpdated": "2026-03-18", - "capabilities": { - "vision": false, - "reasoning": true, - "functionCalling": true - }, - "thinkingOptions": { - "toggle": true - }, - "modalities": { - "input": ["text"], - "output": ["text"] - } }, - "mimo-v2-tts": { - "displayName": "MiMo-V2-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "lifecycle": "active", - "contextWindow": 8192, - "maxOutputTokens": 8192, - "lastUpdated": "2026-03-18", - "capabilities": { - "vision": false, - "reasoning": false, - "functionCalling": false - }, - "modalities": { - "input": ["text"], - "output": ["audio"] - } - }, - "mimo-v2.5": { - "displayName": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", + "mimo-v2.6-flash": { + "displayName": "MiMo-V2.6-Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } + }, + "mimo-v2.6-pro": { + "displayName": "MiMo-V2.6-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } + } + }, + "xiaomi-token-plan-sgp": { + "mimo-v2-pro": { + "displayName": "MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "lifecycle": "deprecated", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "knowledgeCutoff": "2024-12", + "lastUpdated": "2026-03-18", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, + "mimo-v2-tts": { + "displayName": "MiMo-V2-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "lifecycle": "active", + "contextWindow": 8192, + "maxOutputTokens": 8192, + "lastUpdated": "2026-03-18", + "capabilities": { + "vision": false, + "reasoning": false, + "functionCalling": false + }, + "modalities": { + "input": ["text"], + "output": ["audio"] + } + }, + "mimo-v2.5": { + "displayName": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "knowledgeCutoff": "2024-12", + "lastUpdated": "2026-04-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } + }, + "mimo-v2.5-pro": { + "displayName": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "knowledgeCutoff": "2024-12", + "lastUpdated": "2026-04-22", + "capabilities": { + "vision": false, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text"], + "output": ["text"] + } + }, + "mimo-v2.5-tts": { + "displayName": "MiMo-V2.5-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "lifecycle": "active", + "contextWindow": 8192, + "maxOutputTokens": 8192, + "lastUpdated": "2026-04-22", + "capabilities": { + "vision": false, + "reasoning": false, + "functionCalling": false + }, + "modalities": { + "input": ["text"], + "output": ["audio"] + } + }, + "mimo-v2.5-tts-voiceclone": { + "displayName": "MiMo-V2.5-TTS-VoiceClone", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "lifecycle": "active", + "contextWindow": 8192, + "maxOutputTokens": 8192, + "lastUpdated": "2026-04-22", + "capabilities": { + "vision": false, + "reasoning": false, + "functionCalling": false + }, + "modalities": { + "input": ["text"], + "output": ["audio"] + } + }, + "mimo-v2.5-tts-voicedesign": { + "displayName": "MiMo-V2.5-TTS-VoiceDesign", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "lifecycle": "active", + "contextWindow": 8192, + "maxOutputTokens": 8192, + "lastUpdated": "2026-04-22", + "capabilities": { + "vision": false, + "reasoning": false, + "functionCalling": false + }, + "modalities": { + "input": ["text"], + "output": ["audio"] + } + }, + "mimo-v2.6-flash": { + "displayName": "MiMo-V2.6-Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", "lifecycle": "active", "contextWindow": 1048576, "maxOutputTokens": 131072, - "knowledgeCutoff": "2024-12", - "lastUpdated": "2026-04-22", + "lastUpdated": "2026-09-22", "capabilities": { "vision": true, "reasoning": true, @@ -35473,16 +37345,15 @@ "output": ["text"] } }, - "mimo-v2.5-pro": { - "displayName": "MiMo-V2.5-Pro", + "mimo-v2.6-pro": { + "displayName": "MiMo-V2.6-Pro", "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", "lifecycle": "active", "contextWindow": 1048576, "maxOutputTokens": 131072, - "knowledgeCutoff": "2024-12", - "lastUpdated": "2026-04-22", + "lastUpdated": "2026-09-22", "capabilities": { - "vision": false, + "vision": true, "reasoning": true, "functionCalling": true }, @@ -35490,60 +37361,9 @@ "toggle": true }, "modalities": { - "input": ["text"], + "input": ["text", "image", "audio", "video"], "output": ["text"] } - }, - "mimo-v2.5-tts": { - "displayName": "MiMo-V2.5-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "lifecycle": "active", - "contextWindow": 8192, - "maxOutputTokens": 8192, - "lastUpdated": "2026-04-22", - "capabilities": { - "vision": false, - "reasoning": false, - "functionCalling": false - }, - "modalities": { - "input": ["text"], - "output": ["audio"] - } - }, - "mimo-v2.5-tts-voiceclone": { - "displayName": "MiMo-V2.5-TTS-VoiceClone", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "lifecycle": "active", - "contextWindow": 8192, - "maxOutputTokens": 8192, - "lastUpdated": "2026-04-22", - "capabilities": { - "vision": false, - "reasoning": false, - "functionCalling": false - }, - "modalities": { - "input": ["text"], - "output": ["audio"] - } - }, - "mimo-v2.5-tts-voicedesign": { - "displayName": "MiMo-V2.5-TTS-VoiceDesign", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "lifecycle": "active", - "contextWindow": 8192, - "maxOutputTokens": 8192, - "lastUpdated": "2026-04-22", - "capabilities": { - "vision": false, - "reasoning": false, - "functionCalling": false - }, - "modalities": { - "input": ["text"], - "output": ["audio"] - } } }, "xiaomi-token-plan-ams": { @@ -35677,6 +37497,46 @@ "input": ["text"], "output": ["audio"] } + }, + "mimo-v2.6-flash": { + "displayName": "MiMo-V2.6-Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } + }, + "mimo-v2.6-pro": { + "displayName": "MiMo-V2.6-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "lifecycle": "active", + "contextWindow": 1048576, + "maxOutputTokens": 131072, + "lastUpdated": "2026-09-22", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image", "audio", "video"], + "output": ["text"] + } } }, "zai": { @@ -35806,6 +37666,26 @@ "output": ["text"] } }, + "glm-4.6v-flash": { + "displayName": "GLM-4.6V-Flash", + "description": "Lightweight GLM vision model for visual reasoning, documents, and multimodal agents", + "lifecycle": "active", + "contextWindow": 128000, + "maxOutputTokens": 32768, + "lastUpdated": "2025-12-08", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "toggle": true + }, + "modalities": { + "input": ["text", "image", "video"], + "output": ["text"] + } + }, "glm-4.7": { "displayName": "GLM-4.7", "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", @@ -35994,6 +37874,27 @@ "output": ["text"] } }, + "glm-5.3-flashx": { + "displayName": "GLM-5.3-FlashX", + "description": "High-speed GLM-5.3-Flash serving option for coding and agent workflows", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 131072, + "structuredOutput": true, + "lastUpdated": "2026-09-18", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text", "image", "video", "pdf"], + "output": ["text"] + } + }, "glm-5v-turbo": { "displayName": "GLM-5V-Turbo", "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", @@ -37972,117 +39873,39 @@ "output": ["text"] } }, - "x-ai/grok-4": { - "displayName": "Grok 4", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "lifecycle": "deprecated", - "contextWindow": 256000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", - "lastUpdated": "2025-07-09", - "capabilities": { - "vision": true, - "reasoning": true, - "functionCalling": true - }, - "modalities": { - "input": ["image", "text"], - "output": ["text"] - } - }, - "x-ai/grok-4-fast": { - "displayName": "Grok 4 Fast", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "lifecycle": "deprecated", - "contextWindow": 2000000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", - "lastUpdated": "2025-09-19", - "capabilities": { - "vision": true, - "reasoning": true, - "functionCalling": true - }, - "thinkingOptions": { - "toggle": true - }, - "modalities": { - "input": ["text", "image"], - "output": ["text"] - } - }, - "x-ai/grok-4.1-fast": { - "displayName": "Grok 4.1 Fast", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "lifecycle": "deprecated", - "contextWindow": 2000000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", - "lastUpdated": "2025-11-20", - "capabilities": { - "vision": true, - "reasoning": true, - "functionCalling": true - }, - "thinkingOptions": { - "toggle": true - }, - "modalities": { - "input": ["text", "image"], - "output": ["text"] - } - }, - "x-ai/grok-4.1-fast-non-reasoning": { - "displayName": "Grok 4.1 Fast Non Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "lifecycle": "deprecated", - "contextWindow": 2000000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", - "lastUpdated": "2025-11-20", - "capabilities": { - "vision": true, - "reasoning": false, - "functionCalling": true - }, - "modalities": { - "input": ["text", "image"], - "output": ["text"] - } - }, "x-ai/grok-4.2-fast": { "displayName": "Grok 4.2 Fast", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", "lifecycle": "active", "contextWindow": 2000000, "maxOutputTokens": 30000, - "knowledgeCutoff": "2025-08-31", - "lastUpdated": "2026-03-20", + "structuredOutput": true, + "lastUpdated": "2026-03-09", "capabilities": { "vision": true, - "reasoning": true, + "reasoning": false, "functionCalling": true }, "modalities": { - "input": ["text", "image", "video"], + "input": ["text", "image"], "output": ["text"] } }, "x-ai/grok-4.2-fast-non-reasoning": { "displayName": "Grok 4.2 Fast Non Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", "lifecycle": "active", "contextWindow": 2000000, "maxOutputTokens": 30000, - "knowledgeCutoff": "2025-08-31", - "lastUpdated": "2026-03-20", + "structuredOutput": true, + "lastUpdated": "2026-03-09", "capabilities": { "vision": true, "reasoning": false, "functionCalling": true }, "modalities": { - "input": ["text", "image", "video"], + "input": ["text", "image"], "output": ["text"] } }, @@ -38128,6 +39951,28 @@ "output": ["text"] } }, + "x-ai/grok-4.6": { + "displayName": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "lifecycle": "active", + "contextWindow": 500000, + "maxOutputTokens": 500000, + "knowledgeCutoff": "2026-02-01", + "structuredOutput": true, + "lastUpdated": "2026-08-12", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "medium", "high", "xhigh"] + }, + "modalities": { + "input": ["text", "image"], + "output": ["text"] + } + }, "x-ai/grok-build-0.1": { "displayName": "Grok Build 0.1", "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", @@ -38138,32 +39983,64 @@ "lastUpdated": "2026-04-16", "capabilities": { "vision": true, - "reasoning": true, + "reasoning": false, "functionCalling": true }, "modalities": { - "input": ["text", "image", "pdf"], + "input": ["text", "image"], "output": ["text"] } }, - "x-ai/grok-code-fast-1": { - "displayName": "Grok Code Fast 1", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "lifecycle": "deprecated", - "contextWindow": 256000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", - "lastUpdated": "2025-08-26", + "x-ai/grok-imagine-image-2.0": { + "displayName": "Grok Imagine Image 2.0", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "lifecycle": "active", + "contextWindow": 66000, + "lastUpdated": "2026-08-07", + "capabilities": { + "vision": true, + "reasoning": false, + "functionCalling": false + }, + "modalities": { + "input": ["text", "image"], + "output": ["image"] + } + }, + "x-ai/grok-voice-stt-1.0": { + "displayName": "Grok Voice STT 1.0", + "description": "Grok Voice STT 1.0 is xAI's speech-to-text model. It supports transcription with word-level timestamps, optional speaker diarization, and multichannel audio.", + "lifecycle": "active", + "contextWindow": 15000, + "maxOutputTokens": 15000, + "lastUpdated": "2026-08-04", "capabilities": { "vision": false, - "reasoning": true, - "functionCalling": true + "reasoning": false, + "functionCalling": false }, "modalities": { - "input": ["text"], + "input": ["audio"], "output": ["text"] } }, + "x-ai/grok-voice-tts-1.0": { + "displayName": "Grok Voice TTS 1.0", + "description": "Convert text into spoken audio with a single API call. The API supports a rich set of expressive voices, inline speech tags for fine-grained delivery control, and output formats from high-fidelity MP3 to telephony-optimized μ-law.", + "lifecycle": "active", + "contextWindow": 15000, + "maxOutputTokens": 15000, + "lastUpdated": "2026-07-31", + "capabilities": { + "vision": false, + "reasoning": false, + "functionCalling": false + }, + "modalities": { + "input": ["text"], + "output": ["audio"] + } + }, "xiaomi/mimo-v2-flash": { "displayName": "MiMo-V2-Flash", "description": "MiMo flash model for fast multimodal assistance and agent workflows", @@ -38271,12 +40148,13 @@ }, "z-ai/glm-4.5": { "displayName": "GLM 4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", "lifecycle": "active", "contextWindow": 128000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", - "lastUpdated": "2025-07-25", + "maxOutputTokens": 96000, + "knowledgeCutoff": "2025-04", + "structuredOutput": true, + "lastUpdated": "2025-07-28", "capabilities": { "vision": false, "reasoning": true, @@ -38292,12 +40170,12 @@ }, "z-ai/glm-4.5-air": { "displayName": "GLM 4.5 Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", "lifecycle": "active", "contextWindow": 128000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", - "lastUpdated": "2025-07-25", + "maxOutputTokens": 96000, + "knowledgeCutoff": "2025-04", + "lastUpdated": "2025-07-28", "capabilities": { "vision": false, "reasoning": true, @@ -38313,17 +40191,21 @@ }, "z-ai/glm-4.6": { "displayName": "GLM 4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", "lifecycle": "active", "contextWindow": 200000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", + "maxOutputTokens": 128000, + "knowledgeCutoff": "2025-04", + "structuredOutput": true, "lastUpdated": "2025-09-30", "capabilities": { "vision": false, "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "toggle": true + }, "modalities": { "input": ["text"], "output": ["text"] @@ -38334,14 +40216,17 @@ "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "lifecycle": "active", "contextWindow": 200000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", + "maxOutputTokens": 128000, + "knowledgeCutoff": "2025-04", "lastUpdated": "2025-12-08", "capabilities": { "vision": true, "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "toggle": true + }, "modalities": { "input": ["text", "image", "video"], "output": ["text"] @@ -38349,17 +40234,19 @@ }, "z-ai/glm-4.6v-flash": { "displayName": "GLM 4.6V FlashX", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "description": "Lightweight GLM vision model for visual reasoning, documents, and multimodal agents", "lifecycle": "active", "contextWindow": 200000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", + "maxOutputTokens": 128000, "lastUpdated": "2025-12-08", "capabilities": { "vision": true, "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "toggle": true + }, "modalities": { "input": ["text", "image", "video"], "output": ["text"] @@ -38367,35 +40254,41 @@ }, "z-ai/glm-4.6v-flash-free": { "displayName": "GLM 4.6V Flash (Free)", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "description": "Lightweight GLM vision model for visual reasoning, documents, and multimodal agents", "lifecycle": "active", "contextWindow": 200000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", + "maxOutputTokens": 128000, "lastUpdated": "2025-12-08", "capabilities": { "vision": true, "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "toggle": true + }, "modalities": { - "input": ["text", "image", "video"], + "input": ["text", "image", "video", "pdf"], "output": ["text"] } }, "z-ai/glm-4.7": { "displayName": "GLM 4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", "lifecycle": "active", "contextWindow": 200000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", - "lastUpdated": "2025-12-23", + "maxOutputTokens": 128000, + "knowledgeCutoff": "2025-04", + "structuredOutput": true, + "lastUpdated": "2025-12-22", "capabilities": { "vision": false, "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "toggle": true + }, "modalities": { "input": ["text"], "output": ["text"] @@ -38403,17 +40296,20 @@ }, "z-ai/glm-4.7-flash-free": { "displayName": "GLM 4.7 Flash (Free)", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", "lifecycle": "active", "contextWindow": 200000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", + "maxOutputTokens": 128000, + "knowledgeCutoff": "2025-04", "lastUpdated": "2026-01-19", "capabilities": { "vision": false, "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "toggle": true + }, "modalities": { "input": ["text"], "output": ["text"] @@ -38424,14 +40320,17 @@ "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", "lifecycle": "active", "contextWindow": 200000, - "maxOutputTokens": 64000, - "knowledgeCutoff": "2025-01-01", + "maxOutputTokens": 128000, + "knowledgeCutoff": "2025-04", "lastUpdated": "2026-01-19", "capabilities": { "vision": false, "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "toggle": true + }, "modalities": { "input": ["text"], "output": ["text"] @@ -38439,17 +40338,20 @@ }, "z-ai/glm-5": { "displayName": "GLM 5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", "lifecycle": "active", "contextWindow": 200000, "maxOutputTokens": 128000, - "knowledgeCutoff": "2025-01-01", + "structuredOutput": true, "lastUpdated": "2026-02-12", "capabilities": { "vision": false, "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "toggle": true + }, "modalities": { "input": ["text"], "output": ["text"] @@ -38457,30 +40359,33 @@ }, "z-ai/glm-5-turbo": { "displayName": "GLM 5 Turbo", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "description": "Faster GLM-5 lane for coding agents that need lower latency", "lifecycle": "active", "contextWindow": 200000, "maxOutputTokens": 128000, - "knowledgeCutoff": "2025-01-01", - "lastUpdated": "2026-03-20", + "structuredOutput": true, + "lastUpdated": "2026-03-16", "capabilities": { "vision": false, "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "toggle": true + }, "modalities": { "input": ["text"], "output": ["text"] } }, "z-ai/glm-5.1": { - "displayName": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "displayName": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", "lifecycle": "active", "contextWindow": 200000, - "maxOutputTokens": 131072, + "maxOutputTokens": 128000, "structuredOutput": true, - "lastUpdated": "2026-04-03", + "lastUpdated": "2026-04-07", "capabilities": { "vision": false, "reasoning": true, @@ -38499,7 +40404,7 @@ "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "lifecycle": "active", "contextWindow": 1000000, - "maxOutputTokens": 131072, + "maxOutputTokens": 128000, "structuredOutput": true, "lastUpdated": "2026-06-13", "capabilities": { @@ -38515,30 +40420,72 @@ "output": ["text"] } }, - "z-ai/glm-5.2-free": { - "displayName": "GLM 5.2 (Free)", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "z-ai/glm-5.3": { + "displayName": "GLM 5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", "lifecycle": "active", "contextWindow": 1000000, - "maxOutputTokens": 131072, + "maxOutputTokens": 128000, "structuredOutput": true, - "lastUpdated": "2026-06-13", + "lastUpdated": "2026-08-14", "capabilities": { "vision": false, "reasoning": true, "functionCalling": true }, "thinkingOptions": { - "efforts": ["high", "max"] + "efforts": ["low", "high", "max"] }, "modalities": { "input": ["text"], "output": ["text"] } }, + "z-ai/glm-5.3-flash": { + "displayName": "GLM 5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 128000, + "structuredOutput": true, + "lastUpdated": "2026-08-26", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text", "image", "video", "pdf"], + "output": ["text"] + } + }, + "z-ai/glm-5.3-flashx": { + "displayName": "GLM 5.3 FlashX", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "lifecycle": "active", + "contextWindow": 1000000, + "maxOutputTokens": 128000, + "structuredOutput": true, + "lastUpdated": "2026-08-26", + "capabilities": { + "vision": true, + "reasoning": true, + "functionCalling": true + }, + "thinkingOptions": { + "efforts": ["low", "high", "max"] + }, + "modalities": { + "input": ["text", "image", "video", "pdf"], + "output": ["text"] + } + }, "z-ai/glm-5v-turbo": { "displayName": "GLM 5V Turbo", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", "lifecycle": "active", "contextWindow": 200000, "maxOutputTokens": 128000, @@ -38548,10 +40495,29 @@ "reasoning": true, "functionCalling": true }, + "thinkingOptions": { + "toggle": true + }, "modalities": { "input": ["text", "image", "video", "pdf"], "output": ["text"] } + }, + "z-ai/glm-image": { + "displayName": "GLM-Image", + "description": "GLM-Image is an image generation model adopts a hybrid autoregressive + diffusion decoder architecture. In general image generation quality, GLM‑Image aligns with mainstream latent diffusion approaches, but it shows significant advantages in text-rendering and knowledge‑intensive generation scenarios. It performs especially well in tasks requiring precise semantic understanding and complex information expression, while maintaining strong capabilities in high‑fidelity and fine‑grained detail generation. In addition to text‑to‑image generation, GLM‑Image also supports a rich set of image‑to‑image tasks including image editing, style transfer, identity‑preserving generation, and multi‑subject consistency.", + "lifecycle": "active", + "contextWindow": 10240, + "lastUpdated": "2026-01-19", + "capabilities": { + "vision": false, + "reasoning": false, + "functionCalling": false + }, + "modalities": { + "input": ["text"], + "output": ["image"] + } } } }, @@ -38626,6 +40592,13 @@ "cacheReadUsdPer1M": 0.5, "cacheWriteUsdPer1M": 6.25 }, + { + "modelKey": "anthropic:claude-opus-5-5", + "inputUsdPer1M": 4, + "outputUsdPer1M": 20, + "cacheReadUsdPer1M": 0.2, + "cacheWriteUsdPer1M": 5 + }, { "modelKey": "anthropic:claude-sonnet-4-5", "inputUsdPer1M": 3, @@ -38667,6 +40640,12 @@ "cacheReadUsdPer1M": 0.28, "cacheWriteUsdPer1M": 0 }, + { + "modelKey": "alibaba:kimi-k3", + "inputUsdPer1M": 3, + "outputUsdPer1M": 15, + "cacheReadUsdPer1M": 0.3 + }, { "modelKey": "alibaba:qvq-max", "inputUsdPer1M": 1.2, @@ -38988,6 +40967,12 @@ "outputUsdPer1M": 0.87, "cacheReadUsdPer1M": 0.003625 }, + { + "modelKey": "alibaba-cn:deepseek-v4.1-flash", + "inputUsdPer1M": 0.29754, + "outputUsdPer1M": 1.19015, + "cacheReadUsdPer1M": 0.01488 + }, { "modelKey": "alibaba-cn:glm-5.2", "inputUsdPer1M": 1.1, @@ -38995,6 +40980,13 @@ "cacheReadUsdPer1M": 0.275, "cacheWriteUsdPer1M": 0 }, + { + "modelKey": "alibaba-cn:glm-5.3", + "inputUsdPer1M": 1.1, + "outputUsdPer1M": 3.851, + "cacheReadUsdPer1M": 0.275, + "cacheWriteUsdPer1M": 0 + }, { "modelKey": "alibaba-cn:kimi-k2-thinking", "inputUsdPer1M": 0.574, @@ -39010,6 +41002,12 @@ "inputUsdPer1M": 0.929, "outputUsdPer1M": 3.858 }, + { + "modelKey": "alibaba-cn:kimi-k3", + "inputUsdPer1M": 2.827, + "outputUsdPer1M": 14.133, + "cacheReadUsdPer1M": 0.283 + }, { "modelKey": "alibaba-cn:kimi/kimi-k2.5", "inputUsdPer1M": 0.6, @@ -39117,8 +41115,8 @@ }, { "modelKey": "alibaba-cn:qwen-vl-ocr", - "inputUsdPer1M": 0.717, - "outputUsdPer1M": 0.717 + "inputUsdPer1M": 0.043, + "outputUsdPer1M": 0.072 }, { "modelKey": "alibaba-cn:qwen-vl-plus", @@ -39210,16 +41208,6 @@ "inputUsdPer1M": 0.144, "outputUsdPer1M": 0.574 }, - { - "modelKey": "alibaba-cn:qwen3-coder-plus", - "inputUsdPer1M": 1, - "outputUsdPer1M": 5 - }, - { - "modelKey": "alibaba-cn:qwen3-max", - "inputUsdPer1M": 0.861, - "outputUsdPer1M": 3.441 - }, { "modelKey": "alibaba-cn:qwen3-next-80b-a3b-instruct", "inputUsdPer1M": 0.144, @@ -39255,16 +41243,6 @@ "inputUsdPer1M": 0.143353, "outputUsdPer1M": 1.433525 }, - { - "modelKey": "alibaba-cn:qwen3.5-flash", - "inputUsdPer1M": 0.172, - "outputUsdPer1M": 1.72 - }, - { - "modelKey": "alibaba-cn:qwen3.5-plus", - "inputUsdPer1M": 0.573, - "outputUsdPer1M": 3.44 - }, { "modelKey": "alibaba-cn:qwen3.6-flash", "inputUsdPer1M": 0.1875, @@ -39815,9 +41793,9 @@ }, { "modelKey": "deepinfra:tencent/Hy3", - "inputUsdPer1M": 0.14, - "outputUsdPer1M": 0.58, - "cacheReadUsdPer1M": 0.035 + "inputUsdPer1M": 0.13, + "outputUsdPer1M": 0.53, + "cacheReadUsdPer1M": 0.033 }, { "modelKey": "deepinfra:thinkingmachines/Inkling", @@ -39927,6 +41905,12 @@ "outputUsdPer1M": 0.66, "cacheReadUsdPer1M": 0.007 }, + { + "modelKey": "fireworks-ai:accounts/fireworks/models/deepseek-v4-pro", + "inputUsdPer1M": 1.2, + "outputUsdPer1M": 1.2, + "cacheReadUsdPer1M": 0.6 + }, { "modelKey": "fireworks-ai:accounts/fireworks/models/deepseek-v4-pro-0813", "inputUsdPer1M": 1.32, @@ -39987,6 +41971,12 @@ "outputUsdPer1M": 15, "cacheReadUsdPer1M": 0.3 }, + { + "modelKey": "fireworks-ai:accounts/fireworks/models/minimax-m2p7", + "inputUsdPer1M": 1.2, + "outputUsdPer1M": 1.2, + "cacheReadUsdPer1M": 0.6 + }, { "modelKey": "fireworks-ai:accounts/fireworks/models/minimax-m3", "inputUsdPer1M": 0.3, @@ -40003,7 +41993,7 @@ "modelKey": "fireworks-ai:accounts/fireworks/models/nemotron-3-ultra-nvfp4", "inputUsdPer1M": 0.6, "outputUsdPer1M": 2.4, - "cacheReadUsdPer1M": 0.119 + "cacheReadUsdPer1M": 0.12 }, { "modelKey": "fireworks-ai:accounts/fireworks/models/nemotron-lightning-3p5-30b-a3b", @@ -40029,6 +42019,18 @@ "outputUsdPer1M": 6, "cacheReadUsdPer1M": 0.25 }, + { + "modelKey": "fireworks-ai:accounts/fireworks/routers/deepseek-flash-latest", + "inputUsdPer1M": 0.22, + "outputUsdPer1M": 0.66, + "cacheReadUsdPer1M": 0.007 + }, + { + "modelKey": "fireworks-ai:accounts/fireworks/routers/deepseek-pro-latest", + "inputUsdPer1M": 1.32, + "outputUsdPer1M": 3.96, + "cacheReadUsdPer1M": 0.044 + }, { "modelKey": "fireworks-ai:accounts/fireworks/routers/glm-5p2-fast", "inputUsdPer1M": 2.1, @@ -40041,12 +42043,54 @@ "outputUsdPer1M": 6.6, "cacheReadUsdPer1M": 0.39 }, + { + "modelKey": "fireworks-ai:accounts/fireworks/routers/glm-fast-latest", + "inputUsdPer1M": 2.1, + "outputUsdPer1M": 6.6, + "cacheReadUsdPer1M": 0.39 + }, + { + "modelKey": "fireworks-ai:accounts/fireworks/routers/glm-flash-latest", + "inputUsdPer1M": 0.15, + "outputUsdPer1M": 0.5, + "cacheReadUsdPer1M": 0.03 + }, + { + "modelKey": "fireworks-ai:accounts/fireworks/routers/glm-latest", + "inputUsdPer1M": 1.4, + "outputUsdPer1M": 4.4, + "cacheReadUsdPer1M": 0.26 + }, + { + "modelKey": "fireworks-ai:accounts/fireworks/routers/kimi-fast-latest", + "inputUsdPer1M": 4.5, + "outputUsdPer1M": 22.5, + "cacheReadUsdPer1M": 0.45 + }, { "modelKey": "fireworks-ai:accounts/fireworks/routers/kimi-k3-fast", "inputUsdPer1M": 4.5, "outputUsdPer1M": 22.5, "cacheReadUsdPer1M": 0.45 }, + { + "modelKey": "fireworks-ai:accounts/fireworks/routers/kimi-latest", + "inputUsdPer1M": 3, + "outputUsdPer1M": 15, + "cacheReadUsdPer1M": 0.3 + }, + { + "modelKey": "fireworks-ai:accounts/fireworks/routers/minimax-latest", + "inputUsdPer1M": 0.3, + "outputUsdPer1M": 1.2, + "cacheReadUsdPer1M": 0.06 + }, + { + "modelKey": "fireworks-ai:accounts/fireworks/routers/qwen-max-latest", + "inputUsdPer1M": 2, + "outputUsdPer1M": 6, + "cacheReadUsdPer1M": 0.25 + }, { "modelKey": "google:gemini-2.5-flash", "inputUsdPer1M": 0.3, @@ -40559,6 +42603,11 @@ "inputUsdPer1M": 0.14, "outputUsdPer1M": 0.58 }, + { + "modelKey": "huggingface:tencent/Hy4-preview", + "inputUsdPer1M": 0.834, + "outputUsdPer1M": 2.501 + }, { "modelKey": "huggingface:thinkingmachines/Inkling", "inputUsdPer1M": 1, @@ -40848,6 +42897,12 @@ "outputUsdPer1M": 4.4, "cacheReadUsdPer1M": 0.14 }, + { + "modelKey": "mistral:zai-glm-5-3", + "inputUsdPer1M": 1.4, + "outputUsdPer1M": 4.4, + "cacheReadUsdPer1M": 0.14 + }, { "modelKey": "moonshot:kimi-k2.6", "inputUsdPer1M": 0.95, @@ -41414,6 +43469,11 @@ "inputUsdPer1M": 0, "outputUsdPer1M": 0 }, + { + "modelKey": "nvidia:z-ai/glm-5.3", + "inputUsdPer1M": 0, + "outputUsdPer1M": 0 + }, { "modelKey": "nvidia:z-ai/glm-5.3-flash", "inputUsdPer1M": 0, @@ -41825,6 +43885,13 @@ "cacheReadUsdPer1M": 0.5, "cacheWriteUsdPer1M": 6.25 }, + { + "modelKey": "opencode:claude-opus-5-5", + "inputUsdPer1M": 4, + "outputUsdPer1M": 20, + "cacheReadUsdPer1M": 0.2, + "cacheWriteUsdPer1M": 5 + }, { "modelKey": "opencode:claude-sonnet-4-6", "inputUsdPer1M": 3, @@ -41863,6 +43930,12 @@ "outputUsdPer1M": 3.84, "cacheReadUsdPer1M": 0.145 }, + { + "modelKey": "opencode:deepseek-v4.1-flash", + "inputUsdPer1M": 0.3, + "outputUsdPer1M": 1.2, + "cacheReadUsdPer1M": 0.006 + }, { "modelKey": "opencode:gemini-3-flash", "inputUsdPer1M": 0.5, @@ -42168,6 +44241,12 @@ "outputUsdPer1M": 0, "cacheReadUsdPer1M": 0 }, + { + "modelKey": "opencode:mimo-v2.6-flash-free", + "inputUsdPer1M": 0, + "outputUsdPer1M": 0, + "cacheReadUsdPer1M": 0 + }, { "modelKey": "opencode:minimax-m2.1", "inputUsdPer1M": 0.3, @@ -42282,11 +44361,25 @@ "outputUsdPer1M": 0, "cacheReadUsdPer1M": 0 }, + { + "modelKey": "opencode:qwen3.8-flash", + "inputUsdPer1M": 0.15, + "outputUsdPer1M": 0.47, + "cacheReadUsdPer1M": 0.016, + "cacheWriteUsdPer1M": 0.2 + }, { "modelKey": "opencode:ring-2.6-1t-free", "inputUsdPer1M": 0, "outputUsdPer1M": 0 }, + { + "modelKey": "opencode:space-bunny-free", + "inputUsdPer1M": 0, + "outputUsdPer1M": 0, + "cacheReadUsdPer1M": 0, + "cacheWriteUsdPer1M": 0 + }, { "modelKey": "opencode:trinity-large-preview-free", "inputUsdPer1M": 0, @@ -42314,10 +44407,10 @@ }, { "modelKey": "openrouter:~anthropic/claude-opus-latest", - "inputUsdPer1M": 5, - "outputUsdPer1M": 25, - "cacheReadUsdPer1M": 0.5, - "cacheWriteUsdPer1M": 6.25 + "inputUsdPer1M": 4, + "outputUsdPer1M": 20, + "cacheReadUsdPer1M": 0.2, + "cacheWriteUsdPer1M": 5 }, { "modelKey": "openrouter:~anthropic/claude-sonnet-latest", @@ -42328,21 +44421,21 @@ }, { "modelKey": "openrouter:~deepseek/deepseek-flash-latest", - "inputUsdPer1M": 0.15, - "outputUsdPer1M": 0.6, - "cacheReadUsdPer1M": 0.015 + "inputUsdPer1M": 0.1, + "outputUsdPer1M": 0.5, + "cacheReadUsdPer1M": 0.01 }, { "modelKey": "openrouter:~deepseek/deepseek-pro-latest", - "inputUsdPer1M": 0.96, - "outputUsdPer1M": 2.88, - "cacheReadUsdPer1M": 0.088 + "inputUsdPer1M": 0.39996, + "outputUsdPer1M": 1.19988, + "cacheReadUsdPer1M": 0.012726 }, { "modelKey": "openrouter:~deepseek/deepseek-v4-flash-latest", - "inputUsdPer1M": 0.04, - "outputUsdPer1M": 0.1, - "cacheReadUsdPer1M": 0.01 + "inputUsdPer1M": 0.038, + "outputUsdPer1M": 0.55, + "cacheReadUsdPer1M": 0.0228 }, { "modelKey": "openrouter:~google/gemini-flash-latest", @@ -42353,9 +44446,9 @@ }, { "modelKey": "openrouter:~moonshotai/kimi-latest", - "inputUsdPer1M": 2.1, - "outputUsdPer1M": 10.95, - "cacheReadUsdPer1M": 0.23 + "inputUsdPer1M": 1.4989, + "outputUsdPer1M": 10.758, + "cacheReadUsdPer1M": 0.3 }, { "modelKey": "openrouter:~openai/gpt-mini-latest", @@ -42371,9 +44464,9 @@ }, { "modelKey": "openrouter:~z-ai/glm-latest", - "inputUsdPer1M": 0.9, - "outputUsdPer1M": 3, - "cacheReadUsdPer1M": 0.15 + "inputUsdPer1M": 0.5614, + "outputUsdPer1M": 1.7644, + "cacheReadUsdPer1M": 0.10426 }, { "modelKey": "openrouter:aion-labs/aion-2.0", @@ -42393,6 +44486,18 @@ "outputUsdPer1M": 1.4, "cacheReadUsdPer1M": 0.18 }, + { + "modelKey": "openrouter:aion-labs/aion-3.5", + "inputUsdPer1M": 3, + "outputUsdPer1M": 6, + "cacheReadUsdPer1M": 0.75 + }, + { + "modelKey": "openrouter:aion-labs/aion-3.5-mini", + "inputUsdPer1M": 0.7, + "outputUsdPer1M": 1.4, + "cacheReadUsdPer1M": 0.18 + }, { "modelKey": "openrouter:aion-labs/aion-rp-llama-3.1-8b", "inputUsdPer1M": 0.8, @@ -42457,13 +44562,6 @@ "cacheReadUsdPer1M": 0.1, "cacheWriteUsdPer1M": 1.25 }, - { - "modelKey": "openrouter:anthropic/claude-opus-4", - "inputUsdPer1M": 15, - "outputUsdPer1M": 75, - "cacheReadUsdPer1M": 1.5, - "cacheWriteUsdPer1M": 18.75 - }, { "modelKey": "openrouter:anthropic/claude-opus-4.1", "inputUsdPer1M": 15, @@ -42492,6 +44590,13 @@ "cacheReadUsdPer1M": 0.5, "cacheWriteUsdPer1M": 6.25 }, + { + "modelKey": "openrouter:anthropic/claude-opus-5.5", + "inputUsdPer1M": 4, + "outputUsdPer1M": 20, + "cacheReadUsdPer1M": 0.2, + "cacheWriteUsdPer1M": 5 + }, { "modelKey": "openrouter:anthropic/claude-sonnet-5", "inputUsdPer1M": 2, @@ -42531,6 +44636,12 @@ "inputUsdPer1M": 2.5, "outputUsdPer1M": 10 }, + { + "modelKey": "openrouter:cohere/command-a-plus", + "inputUsdPer1M": 0.3, + "outputUsdPer1M": 1.5, + "cacheReadUsdPer1M": 0.15 + }, { "modelKey": "openrouter:cohere/command-r-08-2024", "inputUsdPer1M": 0.15, @@ -42553,8 +44664,8 @@ }, { "modelKey": "openrouter:deepseek/deepseek-chat", - "inputUsdPer1M": 0.2574, - "outputUsdPer1M": 1.0287 + "inputUsdPer1M": 0.32, + "outputUsdPer1M": 0.89 }, { "modelKey": "openrouter:deepseek/deepseek-chat-v3-0324", @@ -42602,15 +44713,15 @@ }, { "modelKey": "openrouter:deepseek/deepseek-v4-flash", - "inputUsdPer1M": 0.088606, - "outputUsdPer1M": 0.177212, - "cacheReadUsdPer1M": 0.017721 + "inputUsdPer1M": 0.08246, + "outputUsdPer1M": 0.16492, + "cacheReadUsdPer1M": 0.016492 }, { "modelKey": "openrouter:deepseek/deepseek-v4-flash-0731", - "inputUsdPer1M": 0.06, - "outputUsdPer1M": 0.12, - "cacheReadUsdPer1M": 0.012 + "inputUsdPer1M": 0.04, + "outputUsdPer1M": 0.64, + "cacheReadUsdPer1M": 0.016 }, { "modelKey": "openrouter:deepseek/deepseek-v4-flash-vision-exp", @@ -42620,21 +44731,21 @@ }, { "modelKey": "openrouter:deepseek/deepseek-v4-pro", - "inputUsdPer1M": 1.6, - "outputUsdPer1M": 3.2, - "cacheReadUsdPer1M": 0.135 + "inputUsdPer1M": 0.953172, + "outputUsdPer1M": 1.906344, + "cacheReadUsdPer1M": 0.079431 }, { "modelKey": "openrouter:deepseek/deepseek-v4-pro-0813", - "inputUsdPer1M": 0.9834, - "outputUsdPer1M": 2.9502, - "cacheReadUsdPer1M": 0.03278 + "inputUsdPer1M": 0.462, + "outputUsdPer1M": 1.386, + "cacheReadUsdPer1M": 0.0154 }, { "modelKey": "openrouter:deepseek/deepseek-v4.1-flash", - "inputUsdPer1M": 0.3, - "outputUsdPer1M": 1.2, - "cacheReadUsdPer1M": 0.006 + "inputUsdPer1M": 0.1, + "outputUsdPer1M": 0.5, + "cacheReadUsdPer1M": 0.01 }, { "modelKey": "openrouter:dots-studio/dots-3-note-preview:free", @@ -42802,8 +44913,8 @@ }, { "modelKey": "openrouter:gryphe/mythomax-l2-13b", - "inputUsdPer1M": 0.06, - "outputUsdPer1M": 0.06 + "inputUsdPer1M": 0.08, + "outputUsdPer1M": 0.11 }, { "modelKey": "openrouter:ibm-granite/granite-4.0-h-micro", @@ -42873,12 +44984,6 @@ "outputUsdPer1M": 0.15, "cacheReadUsdPer1M": 0.03 }, - { - "modelKey": "openrouter:kwaipilot/kat-coder-pro-v2", - "inputUsdPer1M": 0.3, - "outputUsdPer1M": 1.2, - "cacheReadUsdPer1M": 0.06 - }, { "modelKey": "openrouter:kwaipilot/kat-coder-pro-v2.5", "inputUsdPer1M": 0.74, @@ -42944,8 +45049,8 @@ }, { "modelKey": "openrouter:meta/muse-glimmer-30b", - "inputUsdPer1M": 0.35, - "outputUsdPer1M": 1.5, + "inputUsdPer1M": 0.3, + "outputUsdPer1M": 1.2, "cacheReadUsdPer1M": 0.04 }, { @@ -42995,7 +45100,7 @@ }, { "modelKey": "openrouter:minimax/minimax-m1", - "inputUsdPer1M": 0.55, + "inputUsdPer1M": 0.4, "outputUsdPer1M": 2.2 }, { @@ -43075,12 +45180,6 @@ "outputUsdPer1M": 6, "cacheReadUsdPer1M": 0.2 }, - { - "modelKey": "openrouter:mistralai/mistral-large-2512", - "inputUsdPer1M": 0.5, - "outputUsdPer1M": 1.5, - "cacheReadUsdPer1M": 0.05 - }, { "modelKey": "openrouter:mistralai/mistral-medium-3", "inputUsdPer1M": 0.4, @@ -43127,8 +45226,8 @@ }, { "modelKey": "openrouter:mistralai/mistral-small-3.2-24b-instruct", - "inputUsdPer1M": 0.075, - "outputUsdPer1M": 0.2 + "inputUsdPer1M": 0.09375, + "outputUsdPer1M": 0.25 }, { "modelKey": "openrouter:mistralai/mixtral-8x22b-instruct", @@ -43172,15 +45271,15 @@ }, { "modelKey": "openrouter:moonshotai/kimi-k2.7-code", - "inputUsdPer1M": 0.71, - "outputUsdPer1M": 3.5, - "cacheReadUsdPer1M": 0.15 + "inputUsdPer1M": 0.7062, + "outputUsdPer1M": 3.3, + "cacheReadUsdPer1M": 0.18 }, { "modelKey": "openrouter:moonshotai/kimi-k3", - "inputUsdPer1M": 2.648138, - "outputUsdPer1M": 13.282724, - "cacheReadUsdPer1M": 0.302644 + "inputUsdPer1M": 3, + "outputUsdPer1M": 15, + "cacheReadUsdPer1M": 0.3 }, { "modelKey": "openrouter:morph/morph-v3-fast", @@ -43192,11 +45291,23 @@ "inputUsdPer1M": 0.9, "outputUsdPer1M": 1.9 }, + { + "modelKey": "openrouter:nex-agi/nex-n2.5-mini", + "inputUsdPer1M": 0.025, + "outputUsdPer1M": 0.1, + "cacheReadUsdPer1M": 0.0025 + }, { "modelKey": "openrouter:nex-agi/nex-n2.5-mini:free", "inputUsdPer1M": 0, "outputUsdPer1M": 0 }, + { + "modelKey": "openrouter:nex-agi/nex-n2.5-pro", + "inputUsdPer1M": 0.075, + "outputUsdPer1M": 0.25, + "cacheReadUsdPer1M": 0.015 + }, { "modelKey": "openrouter:nex-agi/nex-n2.5-pro:free", "inputUsdPer1M": 0, @@ -43477,14 +45588,14 @@ }, { "modelKey": "openrouter:openai/gpt-oss-120b", - "inputUsdPer1M": 0.037, - "outputUsdPer1M": 0.17 + "inputUsdPer1M": 0.15, + "outputUsdPer1M": 0.6, + "cacheReadUsdPer1M": 0.075 }, { "modelKey": "openrouter:openai/gpt-oss-20b", - "inputUsdPer1M": 0.03, - "outputUsdPer1M": 0.13, - "cacheReadUsdPer1M": 0.03 + "inputUsdPer1M": 0.018, + "outputUsdPer1M": 0.09 }, { "modelKey": "openrouter:openai/gpt-oss-safeguard-20b", @@ -43595,6 +45706,11 @@ "inputUsdPer1M": 0, "outputUsdPer1M": 0 }, + { + "modelKey": "openrouter:prism-ml/ternary-bonsai-2-27b", + "inputUsdPer1M": 0.075, + "outputUsdPer1M": 0.5 + }, { "modelKey": "openrouter:qwen/qwen-2.5-72b-instruct", "inputUsdPer1M": 0.36, @@ -43702,8 +45818,8 @@ }, { "modelKey": "openrouter:qwen/qwen3-vl-30b-a3b-instruct", - "inputUsdPer1M": 0.15, - "outputUsdPer1M": 0.6 + "inputUsdPer1M": 0.13, + "outputUsdPer1M": 0.52 }, { "modelKey": "openrouter:qwen/qwen3-vl-30b-a3b-thinking", @@ -43737,8 +45853,9 @@ }, { "modelKey": "openrouter:qwen/qwen3.5-35b-a3b", - "inputUsdPer1M": 0.1625, - "outputUsdPer1M": 1.3 + "inputUsdPer1M": 0.3125, + "outputUsdPer1M": 1.25, + "cacheReadUsdPer1M": 0.15625 }, { "modelKey": "openrouter:qwen/qwen3.5-397b-a17b", @@ -43758,14 +45875,14 @@ }, { "modelKey": "openrouter:qwen/qwen3.6-27b", - "inputUsdPer1M": 0.3, - "outputUsdPer1M": 2, - "cacheReadUsdPer1M": 0.03 + "inputUsdPer1M": 0.32, + "outputUsdPer1M": 2.7, + "cacheReadUsdPer1M": 0.15 }, { "modelKey": "openrouter:qwen/qwen3.6-35b-a3b", - "inputUsdPer1M": 0.1, - "outputUsdPer1M": 0.9, + "inputUsdPer1M": 0.15, + "outputUsdPer1M": 1, "cacheReadUsdPer1M": 0.05 }, { @@ -43783,9 +45900,14 @@ }, { "modelKey": "openrouter:qwen/qwen3.8-27b", - "inputUsdPer1M": 0.214, - "outputUsdPer1M": 2.55, - "cacheReadUsdPer1M": 0.15 + "inputUsdPer1M": 0.42, + "outputUsdPer1M": 3, + "cacheReadUsdPer1M": 0.085 + }, + { + "modelKey": "openrouter:qwen/qwen3.8-27b:free", + "inputUsdPer1M": 0, + "outputUsdPer1M": 0 }, { "modelKey": "openrouter:qwen/qwen3.8-flash", @@ -43801,6 +45923,12 @@ "cacheReadUsdPer1M": 0.25, "cacheWriteUsdPer1M": 2.5 }, + { + "modelKey": "openrouter:qwen/qwen3.8-omni-flash", + "inputUsdPer1M": 0.15, + "outputUsdPer1M": 0.47, + "cacheReadUsdPer1M": 0.016 + }, { "modelKey": "openrouter:rekaai/reka-edge", "inputUsdPer1M": 0.1, @@ -43936,11 +46064,23 @@ "inputUsdPer1M": 0, "outputUsdPer1M": 0 }, + { + "modelKey": "openrouter:unbiased/pareto", + "inputUsdPer1M": 2.5, + "outputUsdPer1M": 7.5, + "cacheReadUsdPer1M": 0.25 + }, { "modelKey": "openrouter:undi95/remm-slerp-l2-13b", "inputUsdPer1M": 0.35, "outputUsdPer1M": 0.65 }, + { + "modelKey": "openrouter:upstage/solar-mini4", + "inputUsdPer1M": 0.05, + "outputUsdPer1M": 0.2, + "cacheReadUsdPer1M": 0.005 + }, { "modelKey": "openrouter:upstage/solar-pro-3", "inputUsdPer1M": 0.15, @@ -43970,6 +46110,24 @@ "outputUsdPer1M": 0.87, "cacheReadUsdPer1M": 0.0036 }, + { + "modelKey": "openrouter:xiaomi/mimo-v2.6-flash", + "inputUsdPer1M": 0.14, + "outputUsdPer1M": 0.28, + "cacheReadUsdPer1M": 0.0028 + }, + { + "modelKey": "openrouter:xiaomi/mimo-v2.6-pro", + "inputUsdPer1M": 0.435, + "outputUsdPer1M": 0.87, + "cacheReadUsdPer1M": 0.0036 + }, + { + "modelKey": "openrouter:xiaomi/mimo-v2.6-pro-ultraspeed", + "inputUsdPer1M": 4.35, + "outputUsdPer1M": 8.7, + "cacheReadUsdPer1M": 0.036 + }, { "modelKey": "openrouter:z-ai/glm-4.5", "inputUsdPer1M": 0.6, @@ -44031,21 +46189,32 @@ }, { "modelKey": "openrouter:z-ai/glm-5.2", - "inputUsdPer1M": 1.4, - "outputUsdPer1M": 4.4, - "cacheReadUsdPer1M": 0.14 + "inputUsdPer1M": 0.6496, + "outputUsdPer1M": 2.0416, + "cacheReadUsdPer1M": 0.12064 + }, + { + "modelKey": "openrouter:z-ai/glm-5.2:free", + "inputUsdPer1M": 0, + "outputUsdPer1M": 0 }, { "modelKey": "openrouter:z-ai/glm-5.3", - "inputUsdPer1M": 1.4, - "outputUsdPer1M": 4.4, - "cacheReadUsdPer1M": 0.26 + "inputUsdPer1M": 0.84, + "outputUsdPer1M": 2.64, + "cacheReadUsdPer1M": 0.156 }, { "modelKey": "openrouter:z-ai/glm-5.3-flash", "inputUsdPer1M": 0.15, "outputUsdPer1M": 0.5, - "cacheReadUsdPer1M": 0.03 + "cacheReadUsdPer1M": 0.05 + }, + { + "modelKey": "openrouter:z-ai/glm-5.3-flashx", + "inputUsdPer1M": 0.37, + "outputUsdPer1M": 1.25, + "cacheReadUsdPer1M": 0.075 }, { "modelKey": "openrouter:z-ai/glm-5v-turbo", @@ -44053,11 +46222,6 @@ "outputUsdPer1M": 4, "cacheReadUsdPer1M": 0.24 }, - { - "modelKey": "siliconflow:baidu/ERNIE-4.5-300B-A47B", - "inputUsdPer1M": 0.28, - "outputUsdPer1M": 1.1 - }, { "modelKey": "siliconflow:ByteDance-Seed/Seed-OSS-36B-Instruct", "inputUsdPer1M": 0.21, @@ -44086,7 +46250,8 @@ { "modelKey": "siliconflow:deepseek-ai/DeepSeek-V3.2", "inputUsdPer1M": 0.27, - "outputUsdPer1M": 0.42 + "outputUsdPer1M": 0.42, + "cacheReadUsdPer1M": 0.135 }, { "modelKey": "siliconflow:deepseek-ai/DeepSeek-V3.2-Exp", @@ -44095,15 +46260,38 @@ }, { "modelKey": "siliconflow:deepseek-ai/DeepSeek-V4-Flash", - "inputUsdPer1M": 0.14, + "inputUsdPer1M": 0.13, "outputUsdPer1M": 0.28, "cacheReadUsdPer1M": 0.028 }, + { + "modelKey": "siliconflow:deepseek-ai/DeepSeek-V4-Flash-0731", + "inputUsdPer1M": 0.22, + "outputUsdPer1M": 0.66, + "cacheReadUsdPer1M": 0.014 + }, + { + "modelKey": "siliconflow:deepseek-ai/DeepSeek-V4-Flash-Vision-Exp", + "inputUsdPer1M": 0.44, + "outputUsdPer1M": 1.32, + "cacheReadUsdPer1M": 0.028 + }, { "modelKey": "siliconflow:deepseek-ai/DeepSeek-V4-Pro", - "inputUsdPer1M": 1.74, - "outputUsdPer1M": 3.48, - "cacheReadUsdPer1M": 0.145 + "inputUsdPer1M": 1.50162, + "outputUsdPer1M": 3.135, + "cacheReadUsdPer1M": 0.135 + }, + { + "modelKey": "siliconflow:deepseek-ai/DeepSeek-V4-Pro-0813", + "inputUsdPer1M": 1.32, + "outputUsdPer1M": 3.96, + "cacheReadUsdPer1M": 0.044 + }, + { + "modelKey": "siliconflow:google/gemma-4-12B-it", + "inputUsdPer1M": 0.1, + "outputUsdPer1M": 0.3 }, { "modelKey": "siliconflow:google/gemma-4-26B-A4B-it", @@ -44120,12 +46308,24 @@ "inputUsdPer1M": 0.14, "outputUsdPer1M": 0.57 }, + { + "modelKey": "siliconflow:meituan-longcat/LongCat-2.0", + "inputUsdPer1M": 0.75, + "outputUsdPer1M": 2.95, + "cacheReadUsdPer1M": 0.015 + }, { "modelKey": "siliconflow:MiniMaxAI/MiniMax-M2.5", "inputUsdPer1M": 0.3, "outputUsdPer1M": 1.2, "cacheReadUsdPer1M": 0.03 }, + { + "modelKey": "siliconflow:MiniMaxAI/MiniMax-M3", + "inputUsdPer1M": 0.3, + "outputUsdPer1M": 1.2, + "cacheReadUsdPer1M": 0.06 + }, { "modelKey": "siliconflow:moonshotai/Kimi-K2.5", "inputUsdPer1M": 0.45, @@ -44135,8 +46335,26 @@ { "modelKey": "siliconflow:moonshotai/Kimi-K2.6", "inputUsdPer1M": 0.77, - "outputUsdPer1M": 4, - "cacheReadUsdPer1M": 0.2 + "outputUsdPer1M": 3.4, + "cacheReadUsdPer1M": 0.14 + }, + { + "modelKey": "siliconflow:moonshotai/Kimi-K2.7-Code", + "inputUsdPer1M": 0.85916, + "outputUsdPer1M": 3.8, + "cacheReadUsdPer1M": 0.17993 + }, + { + "modelKey": "siliconflow:moonshotai/Kimi-K3", + "inputUsdPer1M": 2.7, + "outputUsdPer1M": 13.5, + "cacheReadUsdPer1M": 0.27 + }, + { + "modelKey": "siliconflow:nex-agi/Nex-N2-Pro", + "inputUsdPer1M": 0.5, + "outputUsdPer1M": 2.5, + "cacheReadUsdPer1M": 0.25 }, { "modelKey": "siliconflow:openai/gpt-oss-120b", @@ -44163,11 +46381,6 @@ "inputUsdPer1M": 0.07, "outputUsdPer1M": 0.28 }, - { - "modelKey": "siliconflow:Qwen/Qwen3-235B-A22B-Thinking-2507", - "inputUsdPer1M": 0.13, - "outputUsdPer1M": 0.6 - }, { "modelKey": "siliconflow:Qwen/Qwen3-30B-A3B-Instruct-2507", "inputUsdPer1M": 0.09, @@ -44193,16 +46406,6 @@ "inputUsdPer1M": 0.25, "outputUsdPer1M": 1 }, - { - "modelKey": "siliconflow:Qwen/Qwen3-VL-235B-A22B-Instruct", - "inputUsdPer1M": 0.3, - "outputUsdPer1M": 1.5 - }, - { - "modelKey": "siliconflow:Qwen/Qwen3-VL-235B-A22B-Thinking", - "inputUsdPer1M": 0.45, - "outputUsdPer1M": 3.5 - }, { "modelKey": "siliconflow:Qwen/Qwen3-VL-30B-A3B-Instruct", "inputUsdPer1M": 0.29, @@ -44263,6 +46466,12 @@ "inputUsdPer1M": 0.2, "outputUsdPer1M": 1.6 }, + { + "modelKey": "siliconflow:Qwen/Qwen3.8-2.4T-A95B", + "inputUsdPer1M": 2, + "outputUsdPer1M": 6, + "cacheReadUsdPer1M": 0.25 + }, { "modelKey": "siliconflow:stepfun-ai/Step-3.5-Flash", "inputUsdPer1M": 0.1, @@ -44274,10 +46483,10 @@ "outputUsdPer1M": 0.57 }, { - "modelKey": "siliconflow:tencent/Hy3-preview", - "inputUsdPer1M": 0.066, - "outputUsdPer1M": 0.26, - "cacheReadUsdPer1M": 0.029 + "modelKey": "siliconflow:tencent/Hy3", + "inputUsdPer1M": 0.132, + "outputUsdPer1M": 0.528, + "cacheReadUsdPer1M": 0.033 }, { "modelKey": "siliconflow:zai-org/GLM-4.5-Air", @@ -44292,18 +46501,32 @@ }, { "modelKey": "siliconflow:zai-org/GLM-5.1", - "inputUsdPer1M": 1.4, - "outputUsdPer1M": 4.4, - "cacheReadUsdPer1M": 0.26, + "inputUsdPer1M": 1.19, + "outputUsdPer1M": 3.74, + "cacheReadUsdPer1M": 0.6, "cacheWriteUsdPer1M": 0 }, { "modelKey": "siliconflow:zai-org/GLM-5.2", + "inputUsdPer1M": 1.302, + "outputUsdPer1M": 4.092, + "cacheReadUsdPer1M": 0.26, + "cacheWriteUsdPer1M": 0 + }, + { + "modelKey": "siliconflow:zai-org/GLM-5.3", "inputUsdPer1M": 1.4, "outputUsdPer1M": 4.4, "cacheReadUsdPer1M": 0.26, "cacheWriteUsdPer1M": 0 }, + { + "modelKey": "siliconflow:zai-org/GLM-5.3-Flash", + "inputUsdPer1M": 0.15, + "outputUsdPer1M": 0.5, + "cacheReadUsdPer1M": 0.03, + "cacheWriteUsdPer1M": 0 + }, { "modelKey": "siliconflow:zai-org/GLM-5V-Turbo", "inputUsdPer1M": 1.2, @@ -44341,6 +46564,12 @@ "outputUsdPer1M": 1.11, "cacheReadUsdPer1M": 0.037 }, + { + "modelKey": "stepfun:step-5-preview", + "inputUsdPer1M": 0.959, + "outputUsdPer1M": 2.741, + "cacheReadUsdPer1M": 0.048 + }, { "modelKey": "stepfun-ai:step-1-32k", "inputUsdPer1M": 2.05, @@ -44371,6 +46600,12 @@ "outputUsdPer1M": 1.11, "cacheReadUsdPer1M": 0.037 }, + { + "modelKey": "stepfun-ai:step-5-preview", + "inputUsdPer1M": 1, + "outputUsdPer1M": 2.7, + "cacheReadUsdPer1M": 0.05 + }, { "modelKey": "togetherai:deepcogito/cogito-v2-1-671b", "inputUsdPer1M": 1.25, @@ -44624,24 +46859,11 @@ "inputUsdPer1M": 0.16, "outputUsdPer1M": 0.64 }, - { - "modelKey": "vercel:alibaba/qwen-3.6-max-preview", - "inputUsdPer1M": 1.3, - "outputUsdPer1M": 7.8, - "cacheReadUsdPer1M": 0.13, - "cacheWriteUsdPer1M": 1.625 - }, { "modelKey": "vercel:alibaba/qwen3-235b-a22b-thinking", "inputUsdPer1M": 0.4, "outputUsdPer1M": 4 }, - { - "modelKey": "vercel:alibaba/qwen3-coder", - "inputUsdPer1M": 1.5, - "outputUsdPer1M": 7.5, - "cacheReadUsdPer1M": 0.3 - }, { "modelKey": "vercel:alibaba/qwen3-coder-30b-a3b", "inputUsdPer1M": 0.15, @@ -44652,30 +46874,6 @@ "inputUsdPer1M": 0.5, "outputUsdPer1M": 1.2 }, - { - "modelKey": "vercel:alibaba/qwen3-coder-plus", - "inputUsdPer1M": 1, - "outputUsdPer1M": 5, - "cacheReadUsdPer1M": 0.2 - }, - { - "modelKey": "vercel:alibaba/qwen3-max", - "inputUsdPer1M": 1.2, - "outputUsdPer1M": 6, - "cacheReadUsdPer1M": 0.24 - }, - { - "modelKey": "vercel:alibaba/qwen3-max-preview", - "inputUsdPer1M": 1.2, - "outputUsdPer1M": 6, - "cacheReadUsdPer1M": 0.24 - }, - { - "modelKey": "vercel:alibaba/qwen3-max-thinking", - "inputUsdPer1M": 1.2, - "outputUsdPer1M": 6, - "cacheReadUsdPer1M": 0.24 - }, { "modelKey": "vercel:alibaba/qwen3-next-80b-a3b-instruct", "inputUsdPer1M": 0.15, @@ -44708,32 +46906,11 @@ "cacheReadUsdPer1M": 0.01, "cacheWriteUsdPer1M": 0.125 }, - { - "modelKey": "vercel:alibaba/qwen3.5-plus", - "inputUsdPer1M": 0.4, - "outputUsdPer1M": 2.5, - "cacheReadUsdPer1M": 0.04, - "cacheWriteUsdPer1M": 0.5 - }, { "modelKey": "vercel:alibaba/qwen3.6-27b", "inputUsdPer1M": 0.6, "outputUsdPer1M": 3.6 }, - { - "modelKey": "vercel:alibaba/qwen3.6-plus", - "inputUsdPer1M": 0.5, - "outputUsdPer1M": 3, - "cacheReadUsdPer1M": 0.05, - "cacheWriteUsdPer1M": 0.625 - }, - { - "modelKey": "vercel:alibaba/qwen3.7-flash", - "inputUsdPer1M": 0.03, - "outputUsdPer1M": 0.13, - "cacheReadUsdPer1M": 0.006, - "cacheWriteUsdPer1M": 0.038 - }, { "modelKey": "vercel:alibaba/qwen3.7-max", "inputUsdPer1M": 2.5, @@ -44741,13 +46918,6 @@ "cacheReadUsdPer1M": 0.5, "cacheWriteUsdPer1M": 3.125 }, - { - "modelKey": "vercel:alibaba/qwen3.7-plus", - "inputUsdPer1M": 0.4, - "outputUsdPer1M": 1.6, - "cacheReadUsdPer1M": 0.08, - "cacheWriteUsdPer1M": 0.5 - }, { "modelKey": "vercel:alibaba/qwen3.8-2.4t-a95b", "inputUsdPer1M": 2, @@ -44772,8 +46942,7 @@ "modelKey": "vercel:alibaba/qwen3.8-max", "inputUsdPer1M": 2, "outputUsdPer1M": 6, - "cacheReadUsdPer1M": 0.25, - "cacheWriteUsdPer1M": 2.5 + "cacheReadUsdPer1M": 0.25 }, { "modelKey": "vercel:alibaba/qwen3.8-max-0902", @@ -44782,6 +46951,12 @@ "cacheReadUsdPer1M": 0.25, "cacheWriteUsdPer1M": 2.5 }, + { + "modelKey": "vercel:alibaba/qwen3.8-omni-flash", + "inputUsdPer1M": 0.15, + "outputUsdPer1M": 0.47, + "cacheReadUsdPer1M": 0.016 + }, { "modelKey": "vercel:amazon/nova-2-lite", "inputUsdPer1M": 0.3, @@ -44894,11 +47069,18 @@ "cacheWriteUsdPer1M": 12.5 }, { - "modelKey": "vercel:anthropic/claude-sonnet-4", - "inputUsdPer1M": 3, - "outputUsdPer1M": 15, - "cacheReadUsdPer1M": 0.3, - "cacheWriteUsdPer1M": 3.75 + "modelKey": "vercel:anthropic/claude-opus-5.5", + "inputUsdPer1M": 4, + "outputUsdPer1M": 20, + "cacheReadUsdPer1M": 0.2, + "cacheWriteUsdPer1M": 5 + }, + { + "modelKey": "vercel:anthropic/claude-opus-5.5-fast", + "inputUsdPer1M": 8, + "outputUsdPer1M": 40, + "cacheReadUsdPer1M": 0.4, + "cacheWriteUsdPer1M": 10 }, { "modelKey": "vercel:anthropic/claude-sonnet-4.5", @@ -44919,18 +47101,6 @@ "inputUsdPer1M": 0.25, "outputUsdPer1M": 0.8999999999999999 }, - { - "modelKey": "vercel:bytedance/seed-1.6", - "inputUsdPer1M": 0.25, - "outputUsdPer1M": 2, - "cacheReadUsdPer1M": 0.05 - }, - { - "modelKey": "vercel:bytedance/seed-1.8", - "inputUsdPer1M": 0.25, - "outputUsdPer1M": 2, - "cacheReadUsdPer1M": 0.05 - }, { "modelKey": "vercel:bytedance/seed-2.1-turbo", "inputUsdPer1M": 0.5, @@ -45001,9 +47171,9 @@ }, { "modelKey": "vercel:deepseek/deepseek-v4.1-flash", - "inputUsdPer1M": 0.15, - "outputUsdPer1M": 0.6, - "cacheReadUsdPer1M": 0.015 + "inputUsdPer1M": 0.3, + "outputUsdPer1M": 1.2, + "cacheReadUsdPer1M": 0.007 }, { "modelKey": "vercel:google/gemini-2.5-flash", @@ -45059,12 +47229,6 @@ "outputUsdPer1M": 1.5, "cacheReadUsdPer1M": 0.03 }, - { - "modelKey": "vercel:google/gemini-3.1-pro-preview", - "inputUsdPer1M": 2, - "outputUsdPer1M": 12, - "cacheReadUsdPer1M": 0.2 - }, { "modelKey": "vercel:google/gemini-3.5-flash", "inputUsdPer1M": 1.5, @@ -45100,6 +47264,16 @@ "outputUsdPer1M": 3.75, "cacheReadUsdPer1M": 0.075 }, + { + "modelKey": "vercel:google/gemini-3.8-live", + "inputUsdPer1M": 0.75, + "outputUsdPer1M": 4.5 + }, + { + "modelKey": "vercel:google/gemini-3.8-live-extended-thinking", + "inputUsdPer1M": 0.75, + "outputUsdPer1M": 4.5 + }, { "modelKey": "vercel:google/gemini-omni-flash-preview", "inputUsdPer1M": 1.5, @@ -45170,33 +47344,21 @@ "outputUsdPer1M": 0 }, { - "modelKey": "vercel:interfaze/interfaze-beta", - "inputUsdPer1M": 1.5, - "outputUsdPer1M": 3.5 + "modelKey": "vercel:inference-net/schematron-v2-small", + "inputUsdPer1M": 0.05, + "outputUsdPer1M": 0.23, + "cacheReadUsdPer1M": 0.05 }, { - "modelKey": "vercel:kwaipilot/kat-coder-air-v2.5", - "inputUsdPer1M": 0.15, - "outputUsdPer1M": 0.6, + "modelKey": "vercel:inference-net/schematron-v2-turbo", + "inputUsdPer1M": 0.03, + "outputUsdPer1M": 0.15, "cacheReadUsdPer1M": 0.03 }, { - "modelKey": "vercel:kwaipilot/kat-coder-pro-v1", - "inputUsdPer1M": 0.3, - "outputUsdPer1M": 1.2, - "cacheReadUsdPer1M": 0.06 - }, - { - "modelKey": "vercel:kwaipilot/kat-coder-pro-v2", - "inputUsdPer1M": 0.3, - "outputUsdPer1M": 1.2, - "cacheReadUsdPer1M": 0.06 - }, - { - "modelKey": "vercel:kwaipilot/kat-coder-pro-v2.5", - "inputUsdPer1M": 0.74, - "outputUsdPer1M": 2.96, - "cacheReadUsdPer1M": 0.15 + "modelKey": "vercel:interfaze/interfaze-beta", + "inputUsdPer1M": 1.5, + "outputUsdPer1M": 3.5 }, { "modelKey": "vercel:meta/llama-3.1-70b", @@ -45319,20 +47481,11 @@ "inputUsdPer1M": 0.3, "outputUsdPer1M": 0.9 }, - { - "modelKey": "vercel:mistral/devstral-2", - "inputUsdPer1M": 0.4, - "outputUsdPer1M": 2 - }, - { - "modelKey": "vercel:mistral/devstral-small-2", - "inputUsdPer1M": 0.1, - "outputUsdPer1M": 0.3 - }, { "modelKey": "vercel:mistral/ministral-14b", "inputUsdPer1M": 0.2, - "outputUsdPer1M": 0.2 + "outputUsdPer1M": 0.2, + "cacheReadUsdPer1M": 0.02 }, { "modelKey": "vercel:mistral/ministral-3b", @@ -45347,32 +47500,31 @@ { "modelKey": "vercel:mistral/mistral-large-3", "inputUsdPer1M": 0.5, - "outputUsdPer1M": 1.5 - }, - { - "modelKey": "vercel:mistral/mistral-medium", - "inputUsdPer1M": 0.4, - "outputUsdPer1M": 2 + "outputUsdPer1M": 1.5, + "cacheReadUsdPer1M": 0.05 }, { "modelKey": "vercel:mistral/mistral-medium-3.5", "inputUsdPer1M": 1.5, - "outputUsdPer1M": 7.5 + "outputUsdPer1M": 7.5, + "cacheReadUsdPer1M": 0.15 }, { "modelKey": "vercel:mistral/mistral-nemo", - "inputUsdPer1M": 0.15, - "outputUsdPer1M": 0.15 + "inputUsdPer1M": 0.04, + "outputUsdPer1M": 0.17 }, { "modelKey": "vercel:mistral/mistral-small", - "inputUsdPer1M": 0.1, - "outputUsdPer1M": 0.3 + "inputUsdPer1M": 0.15, + "outputUsdPer1M": 0.6, + "cacheReadUsdPer1M": 0.015 }, { - "modelKey": "vercel:mistral/pixtral-12b", - "inputUsdPer1M": 0.15, - "outputUsdPer1M": 0.15 + "modelKey": "vercel:mixedbread/toast-1", + "inputUsdPer1M": 0.3, + "outputUsdPer1M": 0.72, + "cacheReadUsdPer1M": 0.036 }, { "modelKey": "vercel:moonshotai/kimi-k2", @@ -45388,8 +47540,7 @@ { "modelKey": "vercel:moonshotai/kimi-k2.5", "inputUsdPer1M": 0.6, - "outputUsdPer1M": 3, - "cacheReadUsdPer1M": 0.1 + "outputUsdPer1M": 3 }, { "modelKey": "vercel:moonshotai/kimi-k2.6", @@ -45650,12 +47801,6 @@ "outputUsdPer1M": 28, "cacheReadUsdPer1M": 0.35 }, - { - "modelKey": "vercel:openai/gpt-5.4", - "inputUsdPer1M": 2.5, - "outputUsdPer1M": 15, - "cacheReadUsdPer1M": 0.25 - }, { "modelKey": "vercel:openai/gpt-5.4-fast", "inputUsdPer1M": 5, @@ -45680,70 +47825,12 @@ "outputUsdPer1M": 1.25, "cacheReadUsdPer1M": 0.02 }, - { - "modelKey": "vercel:openai/gpt-5.4-pro", - "inputUsdPer1M": 30, - "outputUsdPer1M": 180 - }, - { - "modelKey": "vercel:openai/gpt-5.5", - "inputUsdPer1M": 5, - "outputUsdPer1M": 30, - "cacheReadUsdPer1M": 0.5 - }, { "modelKey": "vercel:openai/gpt-5.5-fast", "inputUsdPer1M": 12.5, "outputUsdPer1M": 75, "cacheReadUsdPer1M": 1.25 }, - { - "modelKey": "vercel:openai/gpt-5.5-pro", - "inputUsdPer1M": 30, - "outputUsdPer1M": 180 - }, - { - "modelKey": "vercel:openai/gpt-5.6-luna", - "inputUsdPer1M": 0.2, - "outputUsdPer1M": 1.2, - "cacheReadUsdPer1M": 0.02, - "cacheWriteUsdPer1M": 0.25 - }, - { - "modelKey": "vercel:openai/gpt-5.6-luna-fast", - "inputUsdPer1M": 0.4, - "outputUsdPer1M": 2.4, - "cacheReadUsdPer1M": 0.04, - "cacheWriteUsdPer1M": 0.5 - }, - { - "modelKey": "vercel:openai/gpt-5.6-sol", - "inputUsdPer1M": 2, - "outputUsdPer1M": 10, - "cacheReadUsdPer1M": 0.2, - "cacheWriteUsdPer1M": 2.5 - }, - { - "modelKey": "vercel:openai/gpt-5.6-sol-fast", - "inputUsdPer1M": 4, - "outputUsdPer1M": 20, - "cacheReadUsdPer1M": 0.4, - "cacheWriteUsdPer1M": 5 - }, - { - "modelKey": "vercel:openai/gpt-5.6-terra", - "inputUsdPer1M": 2, - "outputUsdPer1M": 12, - "cacheReadUsdPer1M": 0.2, - "cacheWriteUsdPer1M": 2.5 - }, - { - "modelKey": "vercel:openai/gpt-5.6-terra-fast", - "inputUsdPer1M": 4, - "outputUsdPer1M": 24, - "cacheReadUsdPer1M": 0.4, - "cacheWriteUsdPer1M": 5 - }, { "modelKey": "vercel:openai/gpt-image-1", "inputUsdPer1M": 5, @@ -45783,12 +47870,13 @@ { "modelKey": "vercel:openai/gpt-oss-120b", "inputUsdPer1M": 0.1, - "outputUsdPer1M": 0.5 + "outputUsdPer1M": 0.5, + "cacheReadUsdPer1M": 0.1 }, { "modelKey": "vercel:openai/gpt-oss-20b", - "inputUsdPer1M": 0.05, - "outputUsdPer1M": 0.2 + "inputUsdPer1M": 0.03, + "outputUsdPer1M": 0.14 }, { "modelKey": "vercel:openai/gpt-oss-safeguard-120b", @@ -45877,22 +47965,24 @@ "outputUsdPer1M": 0 }, { - "modelKey": "vercel:sakana/fugu-max", - "inputUsdPer1M": 2, - "outputUsdPer1M": 6, - "cacheReadUsdPer1M": 0.25 + "modelKey": "vercel:quiverai/arrow-2", + "inputUsdPer1M": 4, + "outputUsdPer1M": 20, + "cacheReadUsdPer1M": 0.4, + "cacheWriteUsdPer1M": 5 }, { - "modelKey": "vercel:sakana/fugu-ultra", - "inputUsdPer1M": 5, + "modelKey": "vercel:quiverai/arrow-2-telos", + "inputUsdPer1M": 6, "outputUsdPer1M": 30, - "cacheReadUsdPer1M": 0.5 + "cacheReadUsdPer1M": 0.6, + "cacheWriteUsdPer1M": 7.5 }, { - "modelKey": "vercel:sakana/fugu-ultra-v2", - "inputUsdPer1M": 5, - "outputUsdPer1M": 30, - "cacheReadUsdPer1M": 0.5 + "modelKey": "vercel:sakana/fugu-max", + "inputUsdPer1M": 2, + "outputUsdPer1M": 6, + "cacheReadUsdPer1M": 0.25 }, { "modelKey": "vercel:sakana/namazu", @@ -45912,66 +48002,6 @@ "outputUsdPer1M": 0.5, "cacheReadUsdPer1M": 0.05 }, - { - "modelKey": "vercel:spacexai/grok-4.20-multi-agent", - "inputUsdPer1M": 1.25, - "outputUsdPer1M": 2.5, - "cacheReadUsdPer1M": 0.2 - }, - { - "modelKey": "vercel:spacexai/grok-4.20-multi-agent-beta", - "inputUsdPer1M": 1.25, - "outputUsdPer1M": 2.5, - "cacheReadUsdPer1M": 0.2 - }, - { - "modelKey": "vercel:spacexai/grok-4.20-non-reasoning", - "inputUsdPer1M": 1.25, - "outputUsdPer1M": 2.5, - "cacheReadUsdPer1M": 0.2 - }, - { - "modelKey": "vercel:spacexai/grok-4.20-non-reasoning-beta", - "inputUsdPer1M": 1.25, - "outputUsdPer1M": 2.5, - "cacheReadUsdPer1M": 0.4 - }, - { - "modelKey": "vercel:spacexai/grok-4.20-reasoning", - "inputUsdPer1M": 1.25, - "outputUsdPer1M": 2.5, - "cacheReadUsdPer1M": 0.2 - }, - { - "modelKey": "vercel:spacexai/grok-4.20-reasoning-beta", - "inputUsdPer1M": 1.25, - "outputUsdPer1M": 2.5, - "cacheReadUsdPer1M": 0.2 - }, - { - "modelKey": "vercel:spacexai/grok-4.3", - "inputUsdPer1M": 1.25, - "outputUsdPer1M": 2.5, - "cacheReadUsdPer1M": 0.2 - }, - { - "modelKey": "vercel:spacexai/grok-4.5", - "inputUsdPer1M": 2, - "outputUsdPer1M": 6, - "cacheReadUsdPer1M": 0.3 - }, - { - "modelKey": "vercel:spacexai/grok-4.6", - "inputUsdPer1M": 2, - "outputUsdPer1M": 6, - "cacheReadUsdPer1M": 0.5 - }, - { - "modelKey": "vercel:spacexai/grok-build-0.1", - "inputUsdPer1M": 1, - "outputUsdPer1M": 2, - "cacheReadUsdPer1M": 0.2 - }, { "modelKey": "vercel:stepfun/step-3.5-flash", "inputUsdPer1M": 0.09, @@ -45984,6 +48014,12 @@ "outputUsdPer1M": 1.15, "cacheReadUsdPer1M": 0.04 }, + { + "modelKey": "vercel:stepfun/step-5-preview", + "inputUsdPer1M": 1, + "outputUsdPer1M": 2.7, + "cacheReadUsdPer1M": 0.05 + }, { "modelKey": "vercel:tencent/hy-mt2-lite", "inputUsdPer1M": 0.044, @@ -46035,6 +48071,24 @@ "outputUsdPer1M": 0.87, "cacheReadUsdPer1M": 0.0036 }, + { + "modelKey": "vercel:xiaomi/mimo-v2.6-flash", + "inputUsdPer1M": 0.14, + "outputUsdPer1M": 0.28, + "cacheReadUsdPer1M": 0.0028 + }, + { + "modelKey": "vercel:xiaomi/mimo-v2.6-pro", + "inputUsdPer1M": 0.435, + "outputUsdPer1M": 0.87, + "cacheReadUsdPer1M": 0.0036 + }, + { + "modelKey": "vercel:xiaomi/mimo-v2.6-pro-ultraspeed", + "inputUsdPer1M": 4.35, + "outputUsdPer1M": 8.7, + "cacheReadUsdPer1M": 0.036 + }, { "modelKey": "vercel:zai/glm-4.5", "inputUsdPer1M": 0.6, @@ -46123,6 +48177,12 @@ "outputUsdPer1M": 0.5, "cacheReadUsdPer1M": 0.03 }, + { + "modelKey": "vercel:zai/glm-5.3-flashx", + "inputUsdPer1M": 0.37, + "outputUsdPer1M": 1.25, + "cacheReadUsdPer1M": 0.075 + }, { "modelKey": "vercel:zai/glm-5v-turbo", "inputUsdPer1M": 1.2, @@ -46165,6 +48225,24 @@ "outputUsdPer1M": 2.61, "cacheReadUsdPer1M": 0.0108 }, + { + "modelKey": "xiaomi:mimo-v2.6-flash", + "inputUsdPer1M": 0.14, + "outputUsdPer1M": 0.28, + "cacheReadUsdPer1M": 0.0028 + }, + { + "modelKey": "xiaomi:mimo-v2.6-pro", + "inputUsdPer1M": 0.435, + "outputUsdPer1M": 0.87, + "cacheReadUsdPer1M": 0.0036 + }, + { + "modelKey": "xiaomi:mimo-v2.6-pro-ultraspeed", + "inputUsdPer1M": 4.35, + "outputUsdPer1M": 8.7, + "cacheReadUsdPer1M": 0.036 + }, { "modelKey": "zai:glm-4.5", "inputUsdPer1M": 0.6, @@ -46203,6 +48281,13 @@ "inputUsdPer1M": 0.3, "outputUsdPer1M": 0.9 }, + { + "modelKey": "zai:glm-4.6v-flash", + "inputUsdPer1M": 0, + "outputUsdPer1M": 0, + "cacheReadUsdPer1M": 0, + "cacheWriteUsdPer1M": 0 + }, { "modelKey": "zai:glm-4.7", "inputUsdPer1M": 0.6, @@ -46261,9 +48346,16 @@ }, { "modelKey": "zai:glm-5.3-flash", - "inputUsdPer1M": 0.075, - "outputUsdPer1M": 0.25, - "cacheReadUsdPer1M": 0.015, + "inputUsdPer1M": 0.15, + "outputUsdPer1M": 0.5, + "cacheReadUsdPer1M": 0.03, + "cacheWriteUsdPer1M": 0 + }, + { + "modelKey": "zai:glm-5.3-flashx", + "inputUsdPer1M": 0.37, + "outputUsdPer1M": 1.25, + "cacheReadUsdPer1M": 0.075, "cacheWriteUsdPer1M": 0 }, { @@ -46778,52 +48870,12 @@ "outputUsdPer1M": 1.12, "cacheReadUsdPer1M": 0.03 }, - { - "modelKey": "zenmux:x-ai/grok-4", - "inputUsdPer1M": 3, - "outputUsdPer1M": 15, - "cacheReadUsdPer1M": 0.75 - }, - { - "modelKey": "zenmux:x-ai/grok-4-fast", - "inputUsdPer1M": 0.2, - "outputUsdPer1M": 0.5, - "cacheReadUsdPer1M": 0.05 - }, - { - "modelKey": "zenmux:x-ai/grok-4.1-fast", - "inputUsdPer1M": 0.2, - "outputUsdPer1M": 0.5, - "cacheReadUsdPer1M": 0.05 - }, - { - "modelKey": "zenmux:x-ai/grok-4.1-fast-non-reasoning", - "inputUsdPer1M": 0.2, - "outputUsdPer1M": 0.5, - "cacheReadUsdPer1M": 0.05 - }, - { - "modelKey": "zenmux:x-ai/grok-4.2-fast", - "inputUsdPer1M": 3, - "outputUsdPer1M": 9 - }, - { - "modelKey": "zenmux:x-ai/grok-4.2-fast-non-reasoning", - "inputUsdPer1M": 3, - "outputUsdPer1M": 9 - }, { "modelKey": "zenmux:x-ai/grok-build-0.1", "inputUsdPer1M": 1, "outputUsdPer1M": 2, "cacheReadUsdPer1M": 0.2 }, - { - "modelKey": "zenmux:x-ai/grok-code-fast-1", - "inputUsdPer1M": 0.2, - "outputUsdPer1M": 1.5, - "cacheReadUsdPer1M": 0.02 - }, { "modelKey": "zenmux:xiaomi/mimo-v2-flash", "inputUsdPer1M": 0.1, @@ -46836,92 +48888,41 @@ "outputUsdPer1M": 2, "cacheReadUsdPer1M": 0.08 }, - { - "modelKey": "zenmux:z-ai/glm-4.5", - "inputUsdPer1M": 0.35, - "outputUsdPer1M": 1.54, - "cacheReadUsdPer1M": 0.07 - }, - { - "modelKey": "zenmux:z-ai/glm-4.5-air", - "inputUsdPer1M": 0.11, - "outputUsdPer1M": 0.56, - "cacheReadUsdPer1M": 0.02 - }, - { - "modelKey": "zenmux:z-ai/glm-4.6", - "inputUsdPer1M": 0.35, - "outputUsdPer1M": 1.54, - "cacheReadUsdPer1M": 0.07 - }, - { - "modelKey": "zenmux:z-ai/glm-4.6v", - "inputUsdPer1M": 0.14, - "outputUsdPer1M": 0.42, - "cacheReadUsdPer1M": 0.03 - }, - { - "modelKey": "zenmux:z-ai/glm-4.6v-flash", - "inputUsdPer1M": 0.02, - "outputUsdPer1M": 0.21, - "cacheReadUsdPer1M": 0.0043 - }, - { - "modelKey": "zenmux:z-ai/glm-4.6v-flash-free", - "inputUsdPer1M": 0, - "outputUsdPer1M": 0 - }, - { - "modelKey": "zenmux:z-ai/glm-4.7", - "inputUsdPer1M": 0.28, - "outputUsdPer1M": 1.14, - "cacheReadUsdPer1M": 0.06 - }, { "modelKey": "zenmux:z-ai/glm-4.7-flash-free", "inputUsdPer1M": 0, - "outputUsdPer1M": 0 + "outputUsdPer1M": 0, + "cacheReadUsdPer1M": 0 }, { "modelKey": "zenmux:z-ai/glm-4.7-flashx", - "inputUsdPer1M": 0.07, - "outputUsdPer1M": 0.42, - "cacheReadUsdPer1M": 0.01 + "inputUsdPer1M": 0.0728, + "outputUsdPer1M": 0.4367, + "cacheReadUsdPer1M": 0.0146 }, { - "modelKey": "zenmux:z-ai/glm-5", - "inputUsdPer1M": 0.58, - "outputUsdPer1M": 2.6, - "cacheReadUsdPer1M": 0.14 - }, - { - "modelKey": "zenmux:z-ai/glm-5-turbo", - "inputUsdPer1M": 0.88, - "outputUsdPer1M": 3.48 - }, - { - "modelKey": "zenmux:z-ai/glm-5.1", - "inputUsdPer1M": 0.8781, - "outputUsdPer1M": 3.5126, - "cacheReadUsdPer1M": 0.1903 + "modelKey": "zenmux:z-ai/glm-5.2", + "inputUsdPer1M": 0.98, + "outputUsdPer1M": 3.08, + "cacheReadUsdPer1M": 0.182 }, { - "modelKey": "zenmux:z-ai/glm-5.2", + "modelKey": "zenmux:z-ai/glm-5.3", "inputUsdPer1M": 1.4, - "outputUsdPer1M": 4.5, + "outputUsdPer1M": 4.4, "cacheReadUsdPer1M": 0.26 }, { - "modelKey": "zenmux:z-ai/glm-5.2-free", - "inputUsdPer1M": 0, - "outputUsdPer1M": 0, - "cacheReadUsdPer1M": 0 + "modelKey": "zenmux:z-ai/glm-5.3-flash", + "inputUsdPer1M": 0.15, + "outputUsdPer1M": 0.5, + "cacheReadUsdPer1M": 0.03 }, { - "modelKey": "zenmux:z-ai/glm-5v-turbo", - "inputUsdPer1M": 0.726, - "outputUsdPer1M": 3.1946, - "cacheReadUsdPer1M": 0.1743 + "modelKey": "zenmux:z-ai/glm-5.3-flashx", + "inputUsdPer1M": 0.375, + "outputUsdPer1M": 1.25, + "cacheReadUsdPer1M": 0.075 } ], "providerFacts": { @@ -47022,8 +49023,8 @@ "doc": "https://huggingface.co/docs/inference-providers" }, "kimi-coding-plan": { - "id": "kimi-for-coding", - "name": "Kimi For Coding", + "id": "kimi-code-plan-cn", + "name": "Kimi For Coding (kimi.com)", "api": "https://api.kimi.com/coding/v1", "doc": "https://www.kimi.com/code/docs/en/kimi-code/models.html" }, @@ -47035,8 +49036,8 @@ }, "MiniMax-cn": { "id": "minimax-cn", - "name": "MiniMax (minimaxi.com)", - "api": "https://api.minimaxi.com/anthropic/v1", + "name": "MiniMax (minimax.cn)", + "api": "https://api.minimax.cn/anthropic/v1", "doc": "https://platform.minimaxi.com/docs/guides/quickstart" }, "minimax-coding-plan": { @@ -47089,7 +49090,7 @@ "id": "opencode-go", "name": "OpenCode Go", "api": "https://opencode.ai/zen/go/v1", - "doc": "https://opencode.ai/docs/zen" + "doc": "https://opencode.ai/docs/go" }, "openrouter": { "id": "openrouter", @@ -47267,6 +49268,9 @@ "claude-opus-5": { "npm": "@ai-sdk/anthropic" }, + "claude-opus-5-5": { + "npm": "@ai-sdk/anthropic" + }, "claude-sonnet-4": { "npm": "@ai-sdk/anthropic" }, @@ -47366,12 +49370,21 @@ "gpt-6-astra": { "npm": "@ai-sdk/openai" }, + "gpt-6-luna": { + "npm": "@ai-sdk/openai" + }, + "gpt-6-sol": { + "npm": "@ai-sdk/openai" + }, "grok-4.5": { "npm": "@ai-sdk/openai" }, "grok-4.6": { "npm": "@ai-sdk/openai" }, + "grok-4.7": { + "npm": "@ai-sdk/openai" + }, "grok-build-0.1": { "npm": "@ai-sdk/openai" }, @@ -47404,6 +49417,9 @@ }, "qwen3.6-plus-free": { "npm": "@ai-sdk/anthropic" + }, + "qwen3.8-flash": { + "npm": "@ai-sdk/anthropic" } }, "opencode-go": { @@ -47416,6 +49432,9 @@ "grok-4.6": { "npm": "@ai-sdk/openai" }, + "grok-4.7": { + "npm": "@ai-sdk/openai" + }, "minimax-m2.5": { "npm": "@ai-sdk/anthropic" }, diff --git a/scripts/sync-model-metadata.mjs b/scripts/sync-model-metadata.mjs index 7f65d646f9..c8236d3b3d 100644 --- a/scripts/sync-model-metadata.mjs +++ b/scripts/sync-model-metadata.mjs @@ -566,13 +566,17 @@ function toModelProviderOverride(providerId, modelId, override) { if ( !override || typeof override !== 'object' || - typeof override.npm !== 'string' || + Object.keys(override).some((key) => key !== 'npm' && key !== 'api' && key !== 'body') || + (override.npm !== undefined && typeof override.npm !== 'string') || (override.api !== undefined && typeof override.api !== 'string') ) { throw new Error( `models.dev model ${providerId}/${modelId} has an unsupported provider override`, ); } + // Only `npm` selects a runtime adapter; `body` is a request-body default no + // projection consumes, so an override without `npm` maps to no row. + if (override.npm === undefined) return undefined; return { npm: override.npm, ...(override.api ? { api: override.api } : {}) }; } From 8e01030890705fe0c5642f8f026cc66ff3de1f7b Mon Sep 17 00:00:00 2001 From: AstroHan Date: Thu, 24 Sep 2026 10:23:56 +0800 Subject: [PATCH 2/2] fix(model-thinking): default deepseek-flash to ApplyPatch like its V4 predecessors The registry's DeepSeek fallback list now recommends deepseek-flash first, but APPLY_PATCH_MODELS did not carry the id, so new DeepSeek connections resolved to no ApplyPatch profile while the settings switch still offered the tool. Resolve and test the profile the same way the previous default was pinned. Found by an adversarial review of #5643. Generated-by: Maka (GLM-5.3-Flash) --- packages/core/src/__tests__/model-thinking.test.ts | 1 + packages/core/src/model-thinking.ts | 1 + packages/runtime/src/__tests__/apply-patch-profile.test.ts | 4 ++++ 3 files changed, 6 insertions(+) diff --git a/packages/core/src/__tests__/model-thinking.test.ts b/packages/core/src/__tests__/model-thinking.test.ts index fbc5b8750f..42a9deaa12 100644 --- a/packages/core/src/__tests__/model-thinking.test.ts +++ b/packages/core/src/__tests__/model-thinking.test.ts @@ -253,6 +253,7 @@ test('ApplyPatch defaults are model-specific and explicit choices win', () => { 'gpt-5.6-luna', 'gpt-6-astra', 'gpt-5.4-2026-03-05', + 'deepseek-flash', 'deepseek-v4-flash', 'deepseek-v4-pro', ]) { diff --git a/packages/core/src/model-thinking.ts b/packages/core/src/model-thinking.ts index fc7502a7bd..8f8c8360d6 100644 --- a/packages/core/src/model-thinking.ts +++ b/packages/core/src/model-thinking.ts @@ -163,6 +163,7 @@ const APPLY_PATCH_MODELS: ReadonlySet = new Set([ 'gpt-5.6-terra', 'gpt-5.6-luna', 'gpt-6-astra', + 'deepseek-flash', 'deepseek-v4-flash', 'deepseek-v4-flash-vision-exp', 'deepseek-v4-pro', diff --git a/packages/runtime/src/__tests__/apply-patch-profile.test.ts b/packages/runtime/src/__tests__/apply-patch-profile.test.ts index 1fc96beac3..439da2592a 100644 --- a/packages/runtime/src/__tests__/apply-patch-profile.test.ts +++ b/packages/runtime/src/__tests__/apply-patch-profile.test.ts @@ -33,6 +33,10 @@ describe('ApplyPatch profile routing', () => { resolveModelRuntime({ providerType: 'openai-codex' }, 'gpt-6-astra').applyPatchProfile, { kind: 'codex-v4a-freeform' }, ); + assert.deepEqual( + resolveModelRuntime({ providerType: 'deepseek' }, 'deepseek-flash').applyPatchProfile, + { kind: 'portable-v4a' }, + ); assert.deepEqual( resolveModelRuntime({ providerType: 'deepseek' }, 'deepseek-v4-flash').applyPatchProfile, { kind: 'portable-v4a' },