{ "name": "Van Data Team AI Model Benchmark Dataset", "url": "https://vandatateam.com/ai-model-benchmarks", "citation": "Van Data Team, \"AI Model Benchmarks\", dataset version 2026-10-12, https://vandatateam.com/ai-model-benchmarks", "citationRequest": "Please cite and link the page above when you use or republish this data.", "generatedAt": "2026-10-12", "schemaVersion": "2", "datasetVersion": "2026-10-12", "lastUpdated": "2026-10-12", "scope": { "releaseDateFrom": "2024-01-01", "minModels": 30, "maxModels": 150, "language": "en" }, "providers": [ { "id": "openai", "name": "OpenAI", "website": "https://openai.com/" }, { "id": "anthropic", "name": "Anthropic", "website": "https://www.anthropic.com/" }, { "id": "google", "name": "Google", "website": "https://ai.google.dev/" }, { "id": "meta", "name": "Meta", "website": "https://www.llama.com/" }, { "id": "xai", "name": "xAI", "website": "https://x.ai/" }, { "id": "deepseek", "name": "DeepSeek", "website": "https://www.deepseek.com/" }, { "id": "mistral", "name": "Mistral AI", "website": "https://mistral.ai/" }, { "id": "qwen", "name": "Alibaba Qwen", "website": "https://qwenlm.github.io/" }, { "id": "kimi", "name": "Moonshot Kimi", "website": "https://www.moonshot.ai/" }, { "id": "runway", "name": "Runway", "website": "https://runwayml.com/" }, { "id": "elevenlabs", "name": "ElevenLabs", "website": "https://elevenlabs.io/" }, { "id": "stability-ai", "name": "Stability AI", "website": "https://stability.ai/" }, { "id": "black-forest-labs", "name": "Black Forest Labs", "website": "https://bfl.ai/" } ], "benchmarkDefinitions": [ { "id": "livebench-web-of-lies-v2-2024-11-25", "name": "LiveBench Web of Lies V2", "category": "reasoning", "taskFamily": "text-nlp", "description": "Direct accuracy on the Web of Lies V2 reasoning task.", "unit": "% correct", "direction": "higher-is-better", "minScore": 0, "maxScore": 100, "sourceId": "livebench-2024-11-25-methodology" }, { "id": "livebench-overall-2026-06-25", "name": "LiveBench Overall", "category": "reasoning", "taskFamily": "text-nlp", "description": "Category-weighted overall score across the seven LiveBench categories in the pinned 2026-06-25 snapshot.", "unit": "points", "direction": "higher-is-better", "minScore": 0, "maxScore": 100, "sourceId": "livebench-2026-06-25-averaging-19e766a5" }, { "id": "epoch-gpqa-diamond", "name": "GPQA Diamond", "category": "knowledge", "taskFamily": "text-nlp", "description": "Accuracy on 198 graduate-level multiple-choice science questions, run by Epoch AI. Random guessing scores 25%.", "unit": "percent", "direction": "higher-is-better", "minScore": 0, "maxScore": 100, "sourceId": "epoch-gpqa-diamond" }, { "id": "epoch-frontiermath-tiers-1-3-v2", "name": "FrontierMath Tiers 1-3 (v2)", "category": "reasoning", "taskFamily": "text-nlp", "description": "Share of 295 private research-level math problems solved, run by Epoch AI on the v2 problem set.", "unit": "percent", "direction": "higher-is-better", "minScore": 0, "maxScore": 100, "sourceId": "epoch-frontiermath" }, { "id": "epoch-simpleqa-verified", "name": "SimpleQA Verified", "category": "knowledge", "taskFamily": "text-nlp", "description": "Share of 1,000 short factual questions answered correctly, run by Epoch AI. Higher means fewer confident factual errors.", "unit": "percent", "direction": "higher-is-better", "minScore": 0, "maxScore": 100, "sourceId": "epoch-simpleqa-verified" }, { "id": "arena-text-to-image", "name": "Arena Text-to-Image", "category": "multimodal", "taskFamily": "image", "description": "Human-preference rating from blind pairwise votes on Arena, where people compare images generated from the same prompt and pick the better one. A 100-point gap means the higher-rated model wins about 64% of comparisons.", "unit": "Elo", "direction": "higher-is-better", "minScore": 800, "maxScore": 1600, "sourceId": "arena-text-to-image" }, { "id": "arena-image-edit", "name": "Arena Image Edit", "category": "multimodal", "taskFamily": "image", "description": "Human-preference rating from blind pairwise votes on Arena, where people compare edits of the same image from the same instruction and pick the better one. A 100-point gap means the higher-rated model wins about 64% of comparisons.", "unit": "Elo", "direction": "higher-is-better", "minScore": 800, "maxScore": 1600, "sourceId": "arena-image-edit" }, { "id": "arena-text-to-video", "name": "Arena Text-to-Video", "category": "multimodal", "taskFamily": "video", "description": "Human-preference rating from blind pairwise votes on Arena, where people compare videos generated from the same prompt and pick the better one. A 100-point gap means the higher-rated model wins about 64% of comparisons.", "unit": "Elo", "direction": "higher-is-better", "minScore": 800, "maxScore": 1600, "sourceId": "arena-text-to-video" }, { "id": "arena-image-to-video", "name": "Arena Image-to-Video", "category": "multimodal", "taskFamily": "video", "description": "Human-preference rating from blind pairwise votes on Arena, where people compare videos animated from the same image and prompt and pick the better one. A 100-point gap means the higher-rated model wins about 64% of comparisons.", "unit": "Elo", "direction": "higher-is-better", "minScore": 800, "maxScore": 1600, "sourceId": "arena-image-to-video" }, { "id": "arena-text", "name": "Arena Text", "category": "reasoning", "taskFamily": "text-nlp", "description": "Human-preference rating from blind pairwise votes on Arena, where people compare two anonymous models on the same prompt and pick the better answer. Style-controlled, as on arena.ai. A 100-point gap means the higher-rated model wins about 64% of comparisons.", "unit": "Elo", "direction": "higher-is-better", "minScore": 800, "maxScore": 1600, "sourceId": "arena-text" } ], "benchmarkSnapshots": [ { "id": "livebench-web-of-lies-v2-2024-11-25-snapshot-2024-11-25", "benchmarkId": "livebench-web-of-lies-v2-2024-11-25", "taskFamily": "text-nlp", "version": "2024-11-25", "measuredAt": "2024-11-25", "sourceId": "livebench-2024-11-25-results-347e3d0b" }, { "id": "livebench-overall-2026-06-25-snapshot-2026-06-25", "benchmarkId": "livebench-overall-2026-06-25", "taskFamily": "text-nlp", "version": "2026-06-25", "measuredAt": "2026-06-25", "sourceId": "livebench-2026-06-25-results-19e766a5" }, { "id": "epoch-gpqa-diamond-snapshot-2026-10-12", "benchmarkId": "epoch-gpqa-diamond", "taskFamily": "text-nlp", "version": "Epoch AI benchmark data 2026-10-12", "measuredAt": "2026-10-12", "sourceId": "epoch-benchmark-data-2026-10-12" }, { "id": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "taskFamily": "text-nlp", "version": "Epoch AI benchmark data 2026-10-12", "measuredAt": "2026-10-12", "sourceId": "epoch-benchmark-data-2026-10-12" }, { "id": "epoch-simpleqa-verified-snapshot-2026-10-12", "benchmarkId": "epoch-simpleqa-verified", "taskFamily": "text-nlp", "version": "Epoch AI benchmark data 2026-10-12", "measuredAt": "2026-10-12", "sourceId": "epoch-benchmark-data-2026-10-12" }, { "id": "arena-text-to-image-snapshot-2026-10-07", "benchmarkId": "arena-text-to-image", "taskFamily": "image", "version": "Arena leaderboard published 2026-10-07", "measuredAt": "2026-10-07", "sourceId": "arena-leaderboard-dataset-2026-10-10" }, { "id": "arena-image-edit-snapshot-2026-10-06", "benchmarkId": "arena-image-edit", "taskFamily": "image", "version": "Arena leaderboard published 2026-10-06", "measuredAt": "2026-10-06", "sourceId": "arena-leaderboard-dataset-2026-10-10" }, { "id": "arena-text-to-video-snapshot-2026-09-22", "benchmarkId": "arena-text-to-video", "taskFamily": "video", "version": "Arena leaderboard published 2026-09-22", "measuredAt": "2026-09-22", "sourceId": "arena-leaderboard-dataset-2026-10-10" }, { "id": "arena-image-to-video-snapshot-2026-09-22", "benchmarkId": "arena-image-to-video", "taskFamily": "video", "version": "Arena leaderboard published 2026-09-22", "measuredAt": "2026-09-22", "sourceId": "arena-leaderboard-dataset-2026-10-10" }, { "id": "arena-text-snapshot-2026-10-08", "benchmarkId": "arena-text", "taskFamily": "text-nlp", "version": "Arena leaderboard published 2026-10-08", "measuredAt": "2026-10-08", "sourceId": "arena-leaderboard-dataset-2026-10-10" } ], "models": [ { "id": "gpt-4o", "providerId": "openai", "name": "GPT-4o", "family": "GPT-4o", "releaseDate": "2024-05-13", "status": "legacy", "access": "api", "contextWindowTokens": 128000, "maxOutputTokens": 16384, "modalities": [ "text", "image", "audio" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image", "audio" ], "outputModalities": [ "text" ], "strengths": [ "GPT-4o records API access for text, image, and audio inputs and a published 128,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "API trials combining text, images, and audio" ], "pricing": { "inputPerMillionUsd": 2.5, "outputPerMillionUsd": 10, "cachedInputPerMillionUsd": 1.25, "sourceId": "openai-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 2.5, "tier": "standard", "sourceId": "openai-pricing" }, { "unit": "million-output-tokens", "usd": 10, "tier": "standard", "sourceId": "openai-pricing" }, { "unit": "million-input-tokens", "usd": 1.25, "tier": "cached", "sourceId": "openai-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 48.9, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-4o-2024-05-13" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1346, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gpt-4o-2024-05-13" } ], "officialSourceIds": [ "openai-gpt-4o-release", "openai-pricing", "openai-gpt-4o-model" ] }, { "id": "gpt-41", "providerId": "openai", "name": "GPT-4.1", "family": "GPT-4.1", "releaseDate": "2025-04-14", "status": "active", "access": "api", "contextWindowTokens": 1047576, "maxOutputTokens": 32768, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "GPT-4.1 records API access for text and image inputs and a published 1,047,576-token context window" ], "cautions": [ "Not in the pinned LiveBench snapshot; its independent scores here come from Epoch AI's own GPQA Diamond, FrontierMath and SimpleQA Verified runs" ], "bestFor": [ "Long-context API trials using text and image inputs" ], "pricing": { "inputPerMillionUsd": 2, "outputPerMillionUsd": 8, "cachedInputPerMillionUsd": 0.5, "sourceId": "openai-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 2, "tier": "standard", "sourceId": "openai-pricing" }, { "unit": "million-output-tokens", "usd": 8, "tier": "standard", "sourceId": "openai-pricing" }, { "unit": "million-input-tokens", "usd": 0.5, "tier": "cached", "sourceId": "openai-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 66.9, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-4.1-2025-04-14" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-4.1-2025-04-14" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 31.1, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-4.1-2025-04-14" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1415, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gpt-4.1-2025-04-14" } ], "officialSourceIds": [ "openai-gpt-4-1-release", "openai-pricing", "openai-gpt-4-1-model" ] }, { "id": "gpt-41-mini", "providerId": "openai", "name": "GPT-4.1 mini", "family": "GPT-4.1", "releaseDate": "2025-04-14", "status": "active", "access": "api", "contextWindowTokens": 1047576, "maxOutputTokens": 32768, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "GPT-4.1 mini records API access for text and image inputs and a published 1,047,576-token context window" ], "cautions": [ "Not in the pinned LiveBench snapshot; its independent scores here come from Epoch AI's own GPQA Diamond, FrontierMath and SimpleQA Verified runs" ], "bestFor": [ "Cost-sensitive long-context API evaluation" ], "pricing": { "inputPerMillionUsd": 0.4, "outputPerMillionUsd": 1.6, "cachedInputPerMillionUsd": 0.1, "sourceId": "openai-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.4, "tier": "standard", "sourceId": "openai-pricing" }, { "unit": "million-output-tokens", "usd": 1.6, "tier": "standard", "sourceId": "openai-pricing" }, { "unit": "million-input-tokens", "usd": 0.1, "tier": "cached", "sourceId": "openai-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 65.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-4.1-mini-2025-04-14" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 6.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-4.1-mini-2025-04-14" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 12.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-4.1-mini-2025-04-14" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1383, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gpt-4.1-mini-2025-04-14" } ], "officialSourceIds": [ "openai-gpt-4-1-release", "openai-pricing", "openai-gpt-4-1-mini-model" ] }, { "id": "gpt-41-nano", "providerId": "openai", "name": "GPT-4.1 nano", "family": "GPT-4.1", "releaseDate": "2025-04-14", "status": "active", "access": "api", "contextWindowTokens": 1047576, "maxOutputTokens": 32768, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "GPT-4.1 nano records API access for text and image inputs and a published 1,047,576-token context window" ], "cautions": [ "Not in the pinned LiveBench snapshot; its independent scores here come from Epoch AI's own GPQA Diamond, FrontierMath and SimpleQA Verified runs" ], "bestFor": [ "High-volume API screening at low published token prices" ], "pricing": { "inputPerMillionUsd": 0.1, "outputPerMillionUsd": 0.4, "cachedInputPerMillionUsd": 0.025, "sourceId": "openai-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.1, "tier": "standard", "sourceId": "openai-pricing" }, { "unit": "million-output-tokens", "usd": 0.4, "tier": "standard", "sourceId": "openai-pricing" }, { "unit": "million-input-tokens", "usd": 0.025, "tier": "cached", "sourceId": "openai-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 48.9, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-4.1-nano-2025-04-14" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-4.1-nano-2025-04-14" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1322, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gpt-4.1-nano-2025-04-14" } ], "officialSourceIds": [ "openai-gpt-4-1-release", "openai-pricing", "openai-gpt-4-1-nano-model" ] }, { "id": "claude-3-opus", "providerId": "anthropic", "name": "Claude 3 Opus", "family": "Claude 3", "releaseDate": "2024-03-04", "status": "legacy", "access": "api", "contextWindowTokens": 200000, "maxOutputTokens": 4096, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Claude 3 Opus records API access for text and image inputs and a published 200,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical comparison for systems that still use Claude 3 Opus" ], "pricing": { "inputPerMillionUsd": 15, "outputPerMillionUsd": 75, "cachedInputPerMillionUsd": 1.5, "sourceId": "anthropic-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 15, "tier": "standard", "sourceId": "anthropic-pricing" }, { "unit": "million-output-tokens", "usd": 75, "tier": "standard", "sourceId": "anthropic-pricing" }, { "unit": "million-input-tokens", "usd": 1.5, "tier": "cached", "sourceId": "anthropic-pricing" } ], "benchmarkResults": [ { "benchmarkId": "livebench-web-of-lies-v2-2024-11-25", "benchmarkSnapshotId": "livebench-web-of-lies-v2-2024-11-25-snapshot-2024-11-25", "score": 50, "sourceId": "livebench-2024-11-25-results-347e3d0b", "measuredAt": "2024-11-25", "sourceRow": "claude-3-opus-20240229" }, { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 47.2, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-3-opus-20240229" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 12.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-3-opus-20240229" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1322, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "claude-3-opus-20240229" } ], "officialSourceIds": [ "anthropic-claude-3-release", "anthropic-pricing", "anthropic-models-overview" ] }, { "id": "claude-3-sonnet", "providerId": "anthropic", "name": "Claude 3 Sonnet", "family": "Claude 3", "releaseDate": "2024-03-04", "status": "legacy", "access": "api", "contextWindowTokens": 200000, "maxOutputTokens": 4096, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Claude 3 Sonnet records API access for text and image inputs and a published 200,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical text-and-image API comparison at mid-tier list pricing" ], "pricing": { "inputPerMillionUsd": 3, "outputPerMillionUsd": 15, "cachedInputPerMillionUsd": null, "sourceId": "anthropic-claude-3-release" }, "prices": [ { "unit": "million-input-tokens", "usd": 3, "tier": "standard", "sourceId": "anthropic-claude-3-release" }, { "unit": "million-output-tokens", "usd": 15, "tier": "standard", "sourceId": "anthropic-claude-3-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "anthropic-claude-3-release" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 40.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-3-sonnet-20240229" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1281, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "claude-3-sonnet-20240229" } ], "officialSourceIds": [ "anthropic-claude-3-release", "anthropic-pricing", "anthropic-models-overview" ] }, { "id": "claude-3-haiku", "providerId": "anthropic", "name": "Claude 3 Haiku", "family": "Claude 3", "releaseDate": "2024-03-13", "status": "legacy", "access": "api", "contextWindowTokens": 200000, "maxOutputTokens": 4096, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Claude 3 Haiku records API access for text and image inputs and a published 200,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical cost-sensitive text-and-image API comparison" ], "pricing": { "inputPerMillionUsd": 0.25, "outputPerMillionUsd": 1.25, "cachedInputPerMillionUsd": 0.03, "sourceId": "anthropic-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.25, "tier": "standard", "sourceId": "anthropic-pricing" }, { "unit": "million-output-tokens", "usd": 1.25, "tier": "standard", "sourceId": "anthropic-pricing" }, { "unit": "million-input-tokens", "usd": 0.03, "tier": "cached", "sourceId": "anthropic-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 36.3, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-3-haiku-20240307" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1262, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "claude-3-haiku-20240307" } ], "officialSourceIds": [ "anthropic-claude-3-haiku-release", "anthropic-models-overview", "anthropic-pricing" ] }, { "id": "claude-3-7-sonnet", "providerId": "anthropic", "name": "Claude 3.7 Sonnet", "family": "Claude 3.7", "releaseDate": "2025-02-24", "status": "legacy", "access": "api", "contextWindowTokens": 200000, "maxOutputTokens": 64000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Claude 3.7 Sonnet records API access for text and image inputs and a published 200,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Reasoning screening that may require long generated responses" ], "pricing": { "inputPerMillionUsd": 3, "outputPerMillionUsd": 15, "cachedInputPerMillionUsd": 0.3, "sourceId": "anthropic-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 3, "tier": "standard", "sourceId": "anthropic-pricing" }, { "unit": "million-output-tokens", "usd": 15, "tier": "standard", "sourceId": "anthropic-pricing" }, { "unit": "million-input-tokens", "usd": 0.3, "tier": "cached", "sourceId": "anthropic-pricing" } ], "benchmarkResults": [ { "benchmarkId": "livebench-web-of-lies-v2-2024-11-25", "benchmarkSnapshotId": "livebench-web-of-lies-v2-2024-11-25-snapshot-2024-11-25", "score": 94, "sourceId": "livebench-2024-11-25-results-347e3d0b", "measuredAt": "2024-11-25", "sourceRow": "claude-3-7-sonnet" }, { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 79.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-3-7-sonnet-20250219_64K" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1388, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "claude-3-7-sonnet-20250219-thinking-32k" } ], "officialSourceIds": [ "anthropic-claude-3-7-release", "anthropic-pricing", "anthropic-models-overview" ] }, { "id": "gemini-1-5-pro", "providerId": "google", "name": "Gemini 1.5 Pro", "family": "Gemini 1.5", "releaseDate": "2024-02-15", "status": "legacy", "access": "api", "contextWindowTokens": 2097152, "maxOutputTokens": 8192, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "Gemini 1.5 Pro records API access for text, image, audio, and video inputs and a published 2,097,152-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical very-long-context multimodal API comparison" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 45.9, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-1.5-pro-001" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1324, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gemini-1.5-pro-001" } ], "officialSourceIds": [ "google-gemini-1-5-release", "google-pricing", "google-models-overview" ] }, { "id": "gemini-1-5-flash", "providerId": "google", "name": "Gemini 1.5 Flash", "family": "Gemini 1.5", "releaseDate": "2024-05-14", "status": "legacy", "access": "api", "contextWindowTokens": 1048576, "maxOutputTokens": 8192, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "Gemini 1.5 Flash records API access for text, image, audio, and video inputs and a published 1,048,576-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical multimodal API comparison with long context" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 40.4, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-1.5-flash-001" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1287, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gemini-1.5-flash-001" } ], "officialSourceIds": [ "google-gemini-1-5-flash-release", "google-pricing", "google-models-overview" ] }, { "id": "gemini-2-0-flash", "providerId": "google", "name": "Gemini 2.0 Flash", "family": "Gemini 2.0", "releaseDate": "2024-12-11", "status": "legacy", "access": "api", "contextWindowTokens": 1048576, "maxOutputTokens": 8192, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "Gemini 2.0 Flash records API access for text, image, audio, and video inputs and a published 1,048,576-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical multimodal API reasoning comparison" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" } ], "benchmarkResults": [ { "benchmarkId": "livebench-web-of-lies-v2-2024-11-25", "benchmarkSnapshotId": "livebench-web-of-lies-v2-2024-11-25-snapshot-2024-11-25", "score": 74, "sourceId": "livebench-2024-11-25-results-347e3d0b", "measuredAt": "2024-11-25", "sourceRow": "gemini-2.0-flash" }, { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 64.1, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-2.0-flash-001" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1360, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gemini-2.0-flash-001" } ], "officialSourceIds": [ "google-gemini-2-release", "google-pricing", "google-models-overview" ] }, { "id": "gemini-2-5-pro", "providerId": "google", "name": "Gemini 2.5 Pro", "family": "Gemini 2.5", "releaseDate": "2025-03-25", "status": "active", "access": "api", "contextWindowTokens": 1048576, "maxOutputTokens": 65536, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "Gemini 2.5 Pro records API access for text, image, audio, and video inputs and a published 1,048,576-token context window" ], "cautions": [ "Not in the pinned LiveBench snapshot; its independent scores here come from Epoch AI's own GPQA Diamond, FrontierMath and SimpleQA Verified runs" ], "bestFor": [ "Long-context API trials spanning text, image, audio, and video" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 85.3, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-2.5-pro" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 24.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-2.5-pro" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1445, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gemini-2.5-pro" } ], "officialSourceIds": [ "google-gemini-2-5-release", "google-pricing", "google-gemini-2-5-pro-model" ] }, { "id": "llama-3-8b", "providerId": "meta", "name": "Llama 3 8B", "family": "Llama 3", "releaseDate": "2024-04-18", "status": "legacy", "access": "open-weights", "contextWindowTokens": 8192, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Llama 3 8B records open-weight access for text inputs and a published 8,192-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical small open-weight text deployment baselines" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "meta-llama-3-release" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "meta-llama-3-release" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "meta-llama-3-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "meta-llama-3-release" } ], "benchmarkResults": [ { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1224, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "llama-3-8b-instruct" } ], "officialSourceIds": [ "meta-llama-3-release" ] }, { "id": "llama-3-70b", "providerId": "meta", "name": "Llama 3 70B", "family": "Llama 3", "releaseDate": "2024-04-18", "status": "legacy", "access": "open-weights", "contextWindowTokens": 8192, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Llama 3 70B records open-weight access for text inputs and a published 8,192-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical larger open-weight text deployment baselines" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "meta-llama-3-release" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "meta-llama-3-release" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "meta-llama-3-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "meta-llama-3-release" } ], "benchmarkResults": [ { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1277, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "llama-3-70b-instruct" } ], "officialSourceIds": [ "meta-llama-3-release" ] }, { "id": "llama-3-1-405b", "providerId": "meta", "name": "Llama 3.1 405B", "family": "Llama 3.1", "releaseDate": "2024-07-23", "status": "legacy", "access": "open-weights", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Llama 3.1 405B records open-weight access for text inputs and a published 128,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical large open-weight evaluations with 128,000-token context" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "meta-llama-3-1-release" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "meta-llama-3-1-release" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "meta-llama-3-1-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "meta-llama-3-1-release" } ], "benchmarkResults": [ { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1335, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "llama-3.1-405b-instruct-bf16" } ], "officialSourceIds": [ "meta-llama-3-1-release" ] }, { "id": "llama-3-2-11b-vision", "providerId": "meta", "name": "Llama 3.2 11B Vision", "family": "Llama 3.2", "releaseDate": "2024-09-25", "status": "legacy", "access": "open-weights", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Llama 3.2 11B Vision records open-weight access for text and image inputs and a published 128,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical private text-and-image deployment evaluation" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "meta-llama-3-2-release" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "meta-llama-3-2-release" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "meta-llama-3-2-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "meta-llama-3-2-release" } ], "benchmarkResults": [], "officialSourceIds": [ "meta-llama-3-2-release" ] }, { "id": "grok-1-5", "providerId": "xai", "name": "Grok-1.5", "family": "Grok 1.5", "releaseDate": "2024-03-28", "status": "legacy", "access": "api", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Grok-1.5 records API access for text inputs and a published 128,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical API evaluations needing a documented 128,000-token context" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "xai-models-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "xai-models-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "xai-models-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "xai-models-pricing" } ], "benchmarkResults": [], "officialSourceIds": [ "xai-grok-1-5-release", "xai-models-pricing" ] }, { "id": "grok-1-5v", "providerId": "xai", "name": "Grok-1.5V", "family": "Grok 1.5", "releaseDate": "2024-04-12", "status": "legacy", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Grok-1.5V records API access for text and image inputs" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical text-and-image API comparison" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "xai-models-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "xai-models-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "xai-models-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "xai-models-pricing" } ], "benchmarkResults": [], "officialSourceIds": [ "xai-grok-1-5v-release", "xai-models-pricing" ] }, { "id": "grok-2", "providerId": "xai", "name": "Grok 2", "family": "Grok 2", "releaseDate": "2024-08-13", "status": "legacy", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Grok 2 records API access for text inputs" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical text API comparison where limits can be verified separately" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "xai-models-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "xai-models-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "xai-models-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "xai-models-pricing" } ], "benchmarkResults": [ { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1336, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "grok-2-2024-08-13" } ], "officialSourceIds": [ "xai-grok-2-release", "xai-models-pricing" ] }, { "id": "grok-3", "providerId": "xai", "name": "Grok 3", "family": "Grok 3", "releaseDate": "2025-02-19", "status": "legacy", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Grok 3 records API access for text inputs" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical text API comparison where limits can be verified separately" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "xai-models-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "xai-models-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "xai-models-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "xai-models-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 75.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-3-beta" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1411, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "grok-3-preview-02-24" } ], "officialSourceIds": [ "xai-grok-3-release", "xai-models-pricing", "xai-may-15-retirement" ] }, { "id": "deepseek-v2-chat", "providerId": "deepseek", "name": "DeepSeek-V2 Chat", "family": "DeepSeek-V2", "releaseDate": "2024-05-06", "status": "legacy", "access": "both", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "DeepSeek-V2 Chat records both API and open-weight access for text inputs and a published 128,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical API or open-weight conversational evaluation" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "deepseek-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "deepseek-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "deepseek-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "deepseek-pricing" } ], "benchmarkResults": [], "officialSourceIds": [ "deepseek-v2-release", "deepseek-pricing" ] }, { "id": "deepseek-coder-v2-instruct", "providerId": "deepseek", "name": "DeepSeek-Coder-V2 Instruct", "family": "DeepSeek-Coder-V2", "releaseDate": "2024-06-17", "status": "legacy", "access": "both", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "DeepSeek-Coder-V2 Instruct records both API and open-weight access for text inputs and a published 128,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Repository and code-generation trials with API or open-weight deployment" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "deepseek-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "deepseek-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "deepseek-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "deepseek-pricing" } ], "benchmarkResults": [ { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1265, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "deepseek-coder-v2" } ], "officialSourceIds": [ "deepseek-coder-v2-release", "deepseek-pricing" ] }, { "id": "deepseek-v3", "providerId": "deepseek", "name": "DeepSeek-V3", "family": "DeepSeek-V3", "releaseDate": "2024-12-26", "status": "legacy", "access": "both", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "DeepSeek-V3 records both API and open-weight access for text inputs and a published 128,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical reasoning comparison with API or open-weight access" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "deepseek-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "deepseek-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "deepseek-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "deepseek-pricing" } ], "benchmarkResults": [ { "benchmarkId": "livebench-web-of-lies-v2-2024-11-25", "benchmarkSnapshotId": "livebench-web-of-lies-v2-2024-11-25-snapshot-2024-11-25", "score": 86, "sourceId": "livebench-2024-11-25-results-347e3d0b", "measuredAt": "2024-11-25", "sourceRow": "deepseek-v3" }, { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 56.5, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "DeepSeek-V3" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1358, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "deepseek-v3" } ], "officialSourceIds": [ "deepseek-v3-release", "deepseek-pricing" ] }, { "id": "deepseek-r1", "providerId": "deepseek", "name": "DeepSeek-R1", "family": "DeepSeek-R1", "releaseDate": "2025-01-20", "status": "active", "access": "both", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "DeepSeek-R1 records both API and open-weight access for text inputs and a published 128,000-token context window" ], "cautions": [ "Hosted token price and infrastructure cost are not represented; measure deployment and operating cost directly" ], "bestFor": [ "Reasoning trials requiring API or open-weight deployment options" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "deepseek-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "deepseek-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "deepseek-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "deepseek-pricing" } ], "benchmarkResults": [ { "benchmarkId": "livebench-web-of-lies-v2-2024-11-25", "benchmarkSnapshotId": "livebench-web-of-lies-v2-2024-11-25-snapshot-2024-11-25", "score": 100, "sourceId": "livebench-2024-11-25-results-347e3d0b", "measuredAt": "2024-11-25", "sourceRow": "deepseek-r1" }, { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 71.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "DeepSeek-R1" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1398, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "deepseek-r1" } ], "officialSourceIds": [ "deepseek-r1-release", "deepseek-pricing" ] }, { "id": "codestral-22b", "providerId": "mistral", "name": "Codestral 22B", "family": "Codestral", "releaseDate": "2024-05-29", "status": "legacy", "access": "both", "contextWindowTokens": 32000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Codestral 22B records both API and open-weight access for text inputs and a published 32,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Code-generation trials with API or open-weight deployment" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "mistral-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "mistral-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "mistral-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "mistral-pricing" } ], "benchmarkResults": [], "officialSourceIds": [ "mistral-codestral-release", "mistral-pricing" ] }, { "id": "mistral-large-2", "providerId": "mistral", "name": "Mistral Large 2", "family": "Mistral Large", "releaseDate": "2024-07-24", "status": "legacy", "access": "both", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Mistral Large 2 records both API and open-weight access for text inputs and a published 128,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical general-purpose text trials with flexible deployment" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "mistral-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "mistral-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "mistral-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "mistral-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 49, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "mistral-large-2407" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1314, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "mistral-large-2407" } ], "officialSourceIds": [ "mistral-large-2-release", "mistral-pricing" ] }, { "id": "pixtral-12b", "providerId": "mistral", "name": "Pixtral 12B", "family": "Pixtral", "releaseDate": "2024-09-17", "status": "legacy", "access": "both", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Pixtral 12B records both API and open-weight access for text and image inputs and a published 128,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical private text-and-image evaluation" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "mistral-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "mistral-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "mistral-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "mistral-pricing" } ], "benchmarkResults": [], "officialSourceIds": [ "mistral-pixtral-release", "mistral-pricing" ] }, { "id": "ministral-8b", "providerId": "mistral", "name": "Ministral 8B", "family": "Ministral", "releaseDate": "2024-10-16", "status": "legacy", "access": "both", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Ministral 8B records both API and open-weight access for text inputs and a published 128,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical compact open-weight or API text deployments" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "mistral-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "mistral-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "mistral-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "mistral-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 27.1, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "ministral-8b-2410" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1238, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "ministral-8b-2410" } ], "officialSourceIds": [ "mistral-ministral-release", "mistral-pricing" ] }, { "id": "qwen2-7b-instruct", "providerId": "qwen", "name": "Qwen2 7B Instruct", "family": "Qwen2", "releaseDate": "2024-06-07", "status": "legacy", "access": "open-weights", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Qwen2 7B Instruct records open-weight access for text inputs and a published 128,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical compact private text deployments" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "qwen2-release" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "qwen2-release" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "qwen2-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "qwen2-release" } ], "benchmarkResults": [], "officialSourceIds": [ "qwen2-release" ] }, { "id": "qwen2-72b-instruct", "providerId": "qwen", "name": "Qwen2 72B Instruct", "family": "Qwen2", "releaseDate": "2024-06-07", "status": "legacy", "access": "open-weights", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Qwen2 72B Instruct records open-weight access for text inputs and a published 128,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical larger private text deployments" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "qwen2-release" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "qwen2-release" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "qwen2-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "qwen2-release" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 40.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "qwen2-72b-instruct" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1262, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "qwen2-72b-instruct" } ], "officialSourceIds": [ "qwen2-release" ] }, { "id": "qwen2-5-32b-instruct", "providerId": "qwen", "name": "Qwen2.5 32B Instruct", "family": "Qwen2.5", "releaseDate": "2024-09-19", "status": "legacy", "access": "open-weights", "contextWindowTokens": 128000, "maxOutputTokens": 8192, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Qwen2.5 32B Instruct records open-weight access for text inputs and a published 128,000-token context window" ], "cautions": [ "This record is marked legacy; confirm current availability and migration requirements before adoption" ], "bestFor": [ "Historical private text deployments with a published output limit" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "qwen2-5-release" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "qwen2-5-release" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "qwen2-5-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "qwen2-5-release" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 46.1, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "qwen2.5-32b-instruct" } ], "officialSourceIds": [ "qwen2-5-release" ] }, { "id": "qwen2-5-vl-72b-instruct", "providerId": "qwen", "name": "Qwen2.5-VL 72B Instruct", "family": "Qwen2.5-VL", "releaseDate": "2025-01-26", "status": "active", "access": "open-weights", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text", "image", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image", "video" ], "outputModalities": [ "text" ], "strengths": [ "Qwen2.5-VL 72B Instruct records open-weight access for text, image, and video inputs and a published 128,000-token context window" ], "cautions": [ "Hosted token price and infrastructure cost are not represented; measure deployment and operating cost directly" ], "bestFor": [ "Private text, image, and video evaluation" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "qwen2-5-vl-release" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "qwen2-5-vl-release" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "qwen2-5-vl-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "qwen2-5-vl-release" } ], "benchmarkResults": [], "officialSourceIds": [ "qwen2-5-vl-release" ] }, { "id": "kimi-k2-base", "providerId": "kimi", "name": "Kimi K2 Base", "family": "Kimi K2", "releaseDate": "2025-07-11", "status": "active", "access": "open-weights", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Kimi K2 Base records open-weight access for text inputs and a published 128,000-token context window" ], "cautions": [ "Hosted token price and infrastructure cost are not represented; measure deployment and operating cost directly" ], "bestFor": [ "Base-model adaptation with open-weight infrastructure control" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "kimi-k2-release" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "kimi-k2-release" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "kimi-k2-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "kimi-k2-release" } ], "benchmarkResults": [], "officialSourceIds": [ "kimi-k2-release" ] }, { "id": "kimi-vl-a3b-instruct", "providerId": "kimi", "name": "Kimi-VL-A3B-Instruct", "family": "Kimi-VL", "releaseDate": "2025-04-09", "status": "active", "access": "open-weights", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Kimi-VL-A3B-Instruct records open-weight access for text and image inputs and a published 128,000-token context window" ], "cautions": [ "Hosted token price and infrastructure cost are not represented; measure deployment and operating cost directly" ], "bestFor": [ "Private text-and-image evaluation" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "kimi-vl-release" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "kimi-vl-release" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "kimi-vl-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "kimi-vl-release" } ], "benchmarkResults": [], "officialSourceIds": [ "kimi-vl-release" ] }, { "id": "kimi-dev-72b", "providerId": "kimi", "name": "Kimi-Dev-72B", "family": "Kimi-Dev", "releaseDate": "2025-06-17", "status": "active", "access": "open-weights", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Kimi-Dev-72B records open-weight access for text inputs" ], "cautions": [ "Hosted token price and infrastructure cost are not represented; measure deployment and operating cost directly" ], "bestFor": [ "Private coding and repository evaluation" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "kimi-dev-release" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "kimi-dev-release" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "kimi-dev-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "kimi-dev-release" } ], "benchmarkResults": [], "officialSourceIds": [ "kimi-dev-release" ] }, { "id": "kimi-k2-instruct", "providerId": "kimi", "name": "Kimi K2 Instruct", "family": "Kimi K2", "releaseDate": "2025-07-11", "status": "active", "access": "open-weights", "contextWindowTokens": 128000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Kimi K2 Instruct records open-weight access for text inputs and a published 128,000-token context window" ], "cautions": [ "Hosted token price and infrastructure cost are not represented; measure deployment and operating cost directly" ], "bestFor": [ "Private instruction-following evaluation with 128,000-token context" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "kimi-k2-release" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "kimi-k2-release" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "kimi-k2-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "kimi-k2-release" } ], "benchmarkResults": [ { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1418, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "kimi-k2-0711-preview" } ], "officialSourceIds": [ "kimi-k2-release" ] }, { "id": "gpt-54", "providerId": "openai", "name": "GPT-5.4", "family": "GPT-5.4", "releaseDate": "2026-03-05", "status": "active", "access": "api", "contextWindowTokens": 1050000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "GPT-5.4 provides a 1,050,000-token context window and configurable reasoning for complex professional work" ], "cautions": [ "Prompts above the provider long-context threshold use higher token rates; reproduce cost and quality on the target workload" ], "bestFor": [ "Long-context coding and professional-work evaluations" ], "pricing": { "inputPerMillionUsd": 2.5, "outputPerMillionUsd": 15, "cachedInputPerMillionUsd": 0.25, "sourceId": "openai-gpt-5-4-model" }, "prices": [ { "unit": "million-input-tokens", "usd": 2.5, "tier": "standard", "sourceId": "openai-gpt-5-4-model" }, { "unit": "million-output-tokens", "usd": 15, "tier": "standard", "sourceId": "openai-gpt-5-4-model" }, { "unit": "million-input-tokens", "usd": 0.25, "tier": "cached", "sourceId": "openai-gpt-5-4-model" } ], "benchmarkResults": [ { "benchmarkId": "livebench-overall-2026-06-25", "benchmarkSnapshotId": "livebench-overall-2026-06-25-snapshot-2026-06-25", "score": 77.97, "sourceId": "livebench-2026-06-25-results-19e766a5", "measuredAt": "2026-06-25", "sourceRow": "gpt-5.4-xhigh" }, { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 93.3, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.4-2026-03-05_xhigh" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 78.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.4-2026-03-05_xhigh" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 45.1, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.4-2026-03-05_xhigh" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1475, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gpt-5.4-high" } ], "officialSourceIds": [ "openai-gpt-5-4-model" ] }, { "id": "gpt-55", "providerId": "openai", "name": "GPT-5.5", "family": "GPT-5.5", "releaseDate": "2026-04-23", "status": "active", "access": "api", "contextWindowTokens": 1050000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "GPT-5.5 provides a 1,050,000-token context window for complex coding and professional work" ], "cautions": [ "Prompts above the provider long-context threshold use higher token rates; validate the selected reasoning effort against production cost" ], "bestFor": [ "High-capability API evaluations with long context" ], "pricing": { "inputPerMillionUsd": 5, "outputPerMillionUsd": 30, "cachedInputPerMillionUsd": 0.5, "sourceId": "openai-gpt-5-5-model" }, "prices": [ { "unit": "million-input-tokens", "usd": 5, "tier": "standard", "sourceId": "openai-gpt-5-5-model" }, { "unit": "million-output-tokens", "usd": 30, "tier": "standard", "sourceId": "openai-gpt-5-5-model" }, { "unit": "million-input-tokens", "usd": 0.5, "tier": "cached", "sourceId": "openai-gpt-5-5-model" } ], "benchmarkResults": [ { "benchmarkId": "livebench-overall-2026-06-25", "benchmarkSnapshotId": "livebench-overall-2026-06-25-snapshot-2026-06-25", "score": 79.91, "sourceId": "livebench-2026-06-25-results-19e766a5", "measuredAt": "2026-06-25", "sourceRow": "gpt-5.5-xhigh" }, { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 90.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.5_low" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 85.3, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.5_xhigh" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 63, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.5_xhigh" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1482, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gpt-5.5-high" } ], "officialSourceIds": [ "openai-gpt-5-5-release", "openai-gpt-5-5-model" ] }, { "id": "gpt-56-sol", "providerId": "openai", "name": "GPT-5.6 Sol", "family": "GPT-5.6", "releaseDate": "2026-07-09", "status": "active", "access": "api", "contextWindowTokens": 1050000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "GPT-5.6 Sol is OpenAI's GPT-5.6 flagship for complex professional work, with a 1.05M-token context window and 128K output tokens", "Six reasoning-effort levels from none to max, and the full Responses API tool set including computer use, hosted shell, apply patch and MCP", "Epoch AI measured 93.5% on GPQA Diamond and 89.1% on FrontierMath Tiers 1-3, near the top of this dataset" ], "cautions": [ "Not in the pinned LiveBench snapshot; its independent scores here come from Epoch AI's own GPQA Diamond, FrontierMath and SimpleQA Verified runs" ], "bestFor": [ "Frontier capability evaluations for complex API workflows", "Long-horizon coding and research agents that use computer use or MCP tools" ], "pricing": { "inputPerMillionUsd": 4, "outputPerMillionUsd": 20, "cachedInputPerMillionUsd": 0.4, "sourceId": "openai-api-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 4, "tier": "standard", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 20, "tier": "standard", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.4, "tier": "cached", "sourceId": "openai-api-pricing-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 93.5, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.6-sol_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 89.1, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.6-sol_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 69.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.6-sol_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1485, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gpt-5.6-sol-xhigh" } ], "officialSourceIds": [ "openai-gpt-5-6-release", "openai-gpt-5-6-models", "openai-api-pricing-2026-09", "openai-gpt-5-6-sol-model" ] }, { "id": "gpt-56-terra", "providerId": "openai", "name": "GPT-5.6 Terra", "family": "GPT-5.6", "releaseDate": "2026-07-09", "status": "active", "access": "api", "contextWindowTokens": 1050000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "GPT-5.6 Terra is designed for workloads that balance intelligence and cost: $2 input and $12 output per 1M tokens, half of Sol's input price", "Same 1.05M-token context, 128K output, six effort levels and Responses API tool set as Sol", "Epoch AI measured 93.3% on GPQA Diamond, within a few tenths of Sol, but 43.2% on SimpleQA Verified against Sol's 69.7%" ], "cautions": [ "Not in the pinned LiveBench snapshot; its independent scores here come from Epoch AI's own GPQA Diamond, FrontierMath and SimpleQA Verified runs" ], "bestFor": [ "Balanced long-context API evaluations", "Reasoning-heavy work that does not depend on broad factual recall" ], "pricing": { "inputPerMillionUsd": 2, "outputPerMillionUsd": 12, "cachedInputPerMillionUsd": 0.2, "sourceId": "openai-gpt-5-6-models" }, "prices": [ { "unit": "million-input-tokens", "usd": 2, "tier": "standard", "sourceId": "openai-gpt-5-6-models" }, { "unit": "million-output-tokens", "usd": 12, "tier": "standard", "sourceId": "openai-gpt-5-6-models" }, { "unit": "million-input-tokens", "usd": 0.2, "tier": "cached", "sourceId": "openai-gpt-5-6-models" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 93.3, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.6-terra_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 86, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.6-terra_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 43.2, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.6-terra_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1466, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gpt-5.6-terra-xhigh" } ], "officialSourceIds": [ "openai-gpt-5-6-release", "openai-gpt-5-6-models", "openai-gpt-5-6-terra-model" ] }, { "id": "gpt-56-luna", "providerId": "openai", "name": "GPT-5.6 Luna", "family": "GPT-5.6", "releaseDate": "2026-07-09", "status": "active", "access": "api", "contextWindowTokens": 1050000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "GPT-5.6 Luna is optimized for cost-sensitive workloads at $0.20 input and $1.20 output per 1M tokens", "Keeps the 1.05M-token context, 128K output, six effort levels and full tool set of the larger GPT-5.6 tiers", "Epoch AI measured 91.6% on GPQA Diamond and 82.1% on FrontierMath Tiers 1-3, strong for its price" ], "cautions": [ "Not in the pinned LiveBench snapshot; its independent scores here come from Epoch AI's own GPQA Diamond, FrontierMath and SimpleQA Verified runs" ], "bestFor": [ "High-volume API evaluation with published low token prices", "Subagents and batch reasoning where cost per call dominates" ], "pricing": { "inputPerMillionUsd": 0.2, "outputPerMillionUsd": 1.2, "cachedInputPerMillionUsd": 0.02, "sourceId": "openai-gpt-5-6-models" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.2, "tier": "standard", "sourceId": "openai-gpt-5-6-models" }, { "unit": "million-output-tokens", "usd": 1.2, "tier": "standard", "sourceId": "openai-gpt-5-6-models" }, { "unit": "million-input-tokens", "usd": 0.02, "tier": "cached", "sourceId": "openai-gpt-5-6-models" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 91.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.6-luna_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 82.1, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.6-luna_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 41, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-5.6-luna_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1452, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gpt-5.6-luna-xhigh" } ], "officialSourceIds": [ "openai-gpt-5-6-release", "openai-gpt-5-6-models", "openai-gpt-5-6-luna-model" ] }, { "id": "claude-opus-48", "providerId": "anthropic", "name": "Claude Opus 4.8", "family": "Claude Opus", "releaseDate": "2026-05-28", "status": "active", "access": "api", "contextWindowTokens": 1000000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Claude Opus 4.8 is positioned for coding, agentic tasks, and professional work with a published 1,000,000-token context window" ], "cautions": [ "Maximum output length is not represented because the approved source set does not establish one unambiguously" ], "bestFor": [ "Long-running agentic and professional-work evaluations" ], "pricing": { "inputPerMillionUsd": 5, "outputPerMillionUsd": 25, "cachedInputPerMillionUsd": 0.5, "sourceId": "anthropic-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 5, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 25, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.5, "tier": "cached", "sourceId": "anthropic-pricing-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "livebench-overall-2026-06-25", "benchmarkSnapshotId": "livebench-overall-2026-06-25-snapshot-2026-06-25", "score": 78.93, "sourceId": "livebench-2026-06-25-results-19e766a5", "measuredAt": "2026-06-25", "sourceRow": "claude-opus-4-8-xhigh-effort" }, { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 91, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-opus-4-8_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 80, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-opus-4-8_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 53, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-opus-4-8_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1481, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "claude-opus-4-8-high" } ], "officialSourceIds": [ "anthropic-claude-opus-4-8-release", "anthropic-pricing-2026-09" ] }, { "id": "claude-sonnet-5", "providerId": "anthropic", "name": "Claude Sonnet 5", "family": "Claude Sonnet", "releaseDate": "2026-06-30", "status": "active", "access": "api", "contextWindowTokens": 1000000, "maxOutputTokens": 128000, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Claude Sonnet 5 is positioned for agentic execution, tool use, coding, and knowledge work" ], "cautions": [ "The launch price is introductory through 2026-08-31; refresh pricing before evaluating workloads after that date" ], "bestFor": [ "Agentic coding evaluations with current introductory API pricing" ], "pricing": { "inputPerMillionUsd": 2, "outputPerMillionUsd": 10, "cachedInputPerMillionUsd": 0.2, "sourceId": "anthropic-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 2, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 10, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.2, "tier": "cached", "sourceId": "anthropic-pricing-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "livebench-overall-2026-06-25", "benchmarkSnapshotId": "livebench-overall-2026-06-25-snapshot-2026-06-25", "score": 74.85, "sourceId": "livebench-2026-06-25-results-19e766a5", "measuredAt": "2026-06-25", "sourceRow": "claude-sonnet-5-xhigh-effort" }, { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 90.5, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-sonnet-5_xhigh" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 65.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-sonnet-5_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 33.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-sonnet-5_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1461, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "claude-sonnet-5-high" } ], "officialSourceIds": [ "anthropic-claude-sonnet-5-release", "anthropic-pricing-2026-09" ] }, { "id": "claude-fable-5", "providerId": "anthropic", "name": "Claude Fable 5", "family": "Claude Fable", "releaseDate": "2026-06-09", "status": "active", "access": "api", "contextWindowTokens": 1000000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Claude Fable 5 is Anthropic's highest-capability widely released model for demanding reasoning and long-horizon agentic work" ], "cautions": [ "Fable requires 30-day data retention, and classifier-triggered requests fall back to Claude Opus 4.8" ], "bestFor": [ "Long-horizon coding, reasoning, and knowledge-work evaluations" ], "pricing": { "inputPerMillionUsd": 10, "outputPerMillionUsd": 50, "cachedInputPerMillionUsd": 1, "sourceId": "anthropic-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 10, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 50, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 1, "tier": "cached", "sourceId": "anthropic-pricing-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 85.9, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-fable-5_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 87, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-fable-5_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 70.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-fable-5_xhigh" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1504, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "claude-fable-5-high" } ], "officialSourceIds": [ "anthropic-claude-fable-5-release", "anthropic-models-overview-current", "anthropic-pricing-2026-09" ] }, { "id": "gemini-35-flash", "providerId": "google", "name": "Gemini 3.5 Flash", "family": "Gemini 3.5", "releaseDate": "2026-05-19", "status": "active", "access": "api", "contextWindowTokens": 1048576, "maxOutputTokens": 65536, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "Gemini 3.5 Flash supports text, image, audio, and video inputs with a 1,048,576-token input limit" ], "cautions": [ "Exact current token prices are not represented because the approved source set does not establish a single untiered rate" ], "bestFor": [ "Long-context multimodal and agentic evaluations" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "google-gemini-3-5-model" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "google-gemini-3-5-model" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "google-gemini-3-5-model" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-gemini-3-5-model" } ], "benchmarkResults": [ { "benchmarkId": "livebench-overall-2026-06-25", "benchmarkSnapshotId": "livebench-overall-2026-06-25-snapshot-2026-06-25", "score": 74.64, "sourceId": "livebench-2026-06-25-results-19e766a5", "measuredAt": "2026-06-25", "sourceRow": "gemini-3.5-flash-high" }, { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 92.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.5-flash_high" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 62.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.5-flash_high" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 66.2, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.5-flash_high" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1478, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gemini-3.5-flash-high" } ], "officialSourceIds": [ "google-gemini-3-5-changelog", "google-gemini-3-5-model" ] }, { "id": "gemini-36-flash", "providerId": "google", "name": "Gemini 3.6 Flash", "family": "Gemini 3.6", "releaseDate": "2026-07-21", "status": "active", "access": "api", "contextWindowTokens": 1048576, "maxOutputTokens": 65536, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "Gemini 3.6 Flash supports multimodal inputs and a 1,048,576-token input limit for agentic and coding workflows", "Google says it excels at code generation, agentic execution and spatial reasoning, and lists computer use in preview alongside function calling, code execution and search grounding", "Epoch AI measured 94.1% on GPQA Diamond" ], "cautions": [ "Not in the pinned LiveBench snapshot; its independent scores here come from Epoch AI's own GPQA Diamond, FrontierMath and SimpleQA Verified runs" ], "bestFor": [ "Current multimodal agentic evaluations balancing speed and capability" ], "pricing": { "inputPerMillionUsd": 1.5, "outputPerMillionUsd": 7.5, "cachedInputPerMillionUsd": null, "sourceId": "google-gemini-3-6-latest" }, "prices": [ { "unit": "million-input-tokens", "usd": 1.5, "tier": "standard", "sourceId": "google-gemini-3-6-latest" }, { "unit": "million-output-tokens", "usd": 7.5, "tier": "standard", "sourceId": "google-gemini-3-6-latest" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-gemini-3-6-latest" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 94.1, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.6-flash_high" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 58.9, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.6-flash_high" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 66.2, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.6-flash_high" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1482, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gemini-3.6-flash-high" } ], "officialSourceIds": [ "google-gemini-3-6-model", "google-gemini-3-6-latest" ] }, { "id": "gemini-35-flash-lite", "providerId": "google", "name": "Gemini 3.5 Flash-Lite", "family": "Gemini 3.5", "releaseDate": "2026-07-21", "status": "active", "access": "api", "contextWindowTokens": 1048576, "maxOutputTokens": 65536, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "Gemini 3.5 Flash-Lite is optimized for high-throughput, low-cost execution of subagent tasks and document parsing", "Reads text, images, video, audio and PDF in a 1,048,576-token window with 65,536 output tokens, and supports thinking, function calling, code execution, search grounding and caching" ], "cautions": [ "The represented prices are standard API rates; batch, flex, and priority consumption use different rates" ], "bestFor": [ "High-volume extraction, structured parsing, and cost-sensitive subagent trials" ], "pricing": { "inputPerMillionUsd": 0.3, "outputPerMillionUsd": 2.5, "cachedInputPerMillionUsd": 0.03, "sourceId": "google-gemini-3-5-flash-lite-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.3, "tier": "standard", "sourceId": "google-gemini-3-5-flash-lite-pricing" }, { "unit": "million-output-tokens", "usd": 2.5, "tier": "standard", "sourceId": "google-gemini-3-5-flash-lite-pricing" }, { "unit": "million-input-tokens", "usd": 0.03, "tier": "cached", "sourceId": "google-gemini-3-5-flash-lite-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 83.3, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.5-flash-lite_high" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 26, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.5-flash-lite_high" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1455, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gemini-3.5-flash-lite" } ], "officialSourceIds": [ "google-gemini-3-5-flash-lite-model", "google-gemini-3-5-flash-lite-pricing", "google-gemini-3-5-changelog" ] }, { "id": "grok-420", "providerId": "xai", "name": "Grok 4.20", "family": "Grok 4", "releaseDate": "2026-03-10", "status": "active", "access": "api", "contextWindowTokens": 1000000, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Grok 4.20 supports text and image input and is available in reasoning, non-reasoning, and multi-agent API variants" ], "cautions": [ "The published base rates double when requests reach the provider's 200,000-token long-context threshold" ], "bestFor": [ "Comparing single-agent and multi-agent reasoning on long-context workloads" ], "pricing": { "inputPerMillionUsd": 1.25, "outputPerMillionUsd": 2.5, "cachedInputPerMillionUsd": 0.2, "sourceId": "xai-current-model-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 1.25, "tier": "standard", "sourceId": "xai-current-model-pricing" }, { "unit": "million-output-tokens", "usd": 2.5, "tier": "standard", "sourceId": "xai-current-model-pricing" }, { "unit": "million-input-tokens", "usd": 0.2, "tier": "cached", "sourceId": "xai-current-model-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 89.3, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.20-0309-reasoning" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 44.9, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.20-0309-reasoning" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 30.2, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.20-0309-reasoning" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1472, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "grok-4.20-beta-0309-reasoning" } ], "officialSourceIds": [ "xai-release-notes-current", "xai-grok-4-20-system-card", "xai-current-model-pricing" ] }, { "id": "grok-43", "providerId": "xai", "name": "Grok 4.3", "family": "Grok 4", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": 1000000, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Grok 4.3 is xAI's general-purpose default with configurable reasoning and strong tool calling" ], "cautions": [ "The published base rates double when requests reach the provider's 200,000-token long-context threshold" ], "bestFor": [ "General-purpose reasoning, instruction following, and agentic tool-use evaluations" ], "pricing": { "inputPerMillionUsd": 1.25, "outputPerMillionUsd": 2.5, "cachedInputPerMillionUsd": 0.2, "sourceId": "xai-grok-4-3-model" }, "prices": [ { "unit": "million-input-tokens", "usd": 1.25, "tier": "standard", "sourceId": "xai-grok-4-3-model" }, { "unit": "million-output-tokens", "usd": 2.5, "tier": "standard", "sourceId": "xai-grok-4-3-model" }, { "unit": "million-input-tokens", "usd": 0.2, "tier": "cached", "sourceId": "xai-grok-4-3-model" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 88.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.3_high" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 42.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.3_high" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 33.2, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.3_high" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1442, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "grok-4.3" } ], "officialSourceIds": [ "xai-grok-4-3-model", "xai-grok-4-3-public-availability", "xai-inference-models-api", "xai-may-15-retirement" ] }, { "id": "grok-45", "providerId": "xai", "name": "Grok 4.5", "family": "Grok 4", "releaseDate": "2026-07-16", "status": "active", "access": "api", "contextWindowTokens": 500000, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Grok 4.5 is positioned for coding, agentic tasks, and knowledge work with a 500,000-token context window", "xAI says it is served at about 80 tokens per second and solves tasks in under half the steps, at $2 input and $6 output per 1M tokens" ], "cautions": [ "The published base rates increase for requests at or above the provider long-context threshold" ], "bestFor": [ "Coding and agentic API evaluations below the long-context pricing threshold" ], "pricing": { "inputPerMillionUsd": 2, "outputPerMillionUsd": 6, "cachedInputPerMillionUsd": 0.3, "sourceId": "xai-grok-4-5-model" }, "prices": [ { "unit": "million-input-tokens", "usd": 2, "tier": "standard", "sourceId": "xai-grok-4-5-model" }, { "unit": "million-output-tokens", "usd": 6, "tier": "standard", "sourceId": "xai-grok-4-5-model" }, { "unit": "million-input-tokens", "usd": 0.3, "tier": "cached", "sourceId": "xai-grok-4-5-model" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 93.4, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.5_high" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 57.2, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.5_high" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 48.3, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.5_high" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1466, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "grok-4.5" } ], "officialSourceIds": [ "xai-grok-4-5-release", "xai-grok-4-5-model" ] }, { "id": "grok-46", "providerId": "xai", "name": "Grok 4.6", "family": "Grok 4", "releaseDate": "2026-08-12", "status": "active", "access": "api", "contextWindowTokens": 500000, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Grok 4.6 is positioned for coding, agentic tasks, and knowledge work; its API documentation also lists function calling and structured outputs", "xAI says it focuses on long-running agents that stay with complex tasks across many steps, with more self-testing and verification than Grok 4.5", "Epoch AI measured 94.0% on GPQA Diamond and 66.0% on FrontierMath Tiers 1-3" ], "cautions": [ "Published token rates increase for all tokens in a request once its prompt reaches xAI's 200,000-token long-context threshold" ], "bestFor": [ "Coding, agentic, and knowledge-work API evaluations with controlled long-context cost" ], "pricing": { "inputPerMillionUsd": 2, "outputPerMillionUsd": 6, "cachedInputPerMillionUsd": 0.5, "sourceId": "xai-grok-4-6-model" }, "prices": [ { "unit": "million-input-tokens", "usd": 2, "tier": "standard", "sourceId": "xai-grok-4-6-model" }, { "unit": "million-output-tokens", "usd": 6, "tier": "standard", "sourceId": "xai-grok-4-6-model" }, { "unit": "million-input-tokens", "usd": 0.5, "tier": "cached", "sourceId": "xai-grok-4-6-model" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 94, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.6_high" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 66, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.6_xhigh" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 49.3, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.6_high" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1454, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "grok-4.6-high" } ], "officialSourceIds": [ "xai-grok-4-6-release", "xai-grok-4-6-model", "xai-current-model-pricing" ] }, { "id": "gpt-image-2", "providerId": "openai", "name": "GPT Image 2", "family": "GPT Image", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "image" ], "capabilities": [ "image-generation", "image-editing", "image-to-image" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "image" ], "strengths": [ "Official documentation supports image generation, editing, and high-fidelity image inputs" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "API image generation and editing" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "specialist-openai-gpt-image-2" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-openai-gpt-image-2" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-openai-gpt-image-2" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-openai-gpt-image-2" } ], "benchmarkResults": [ { "benchmarkId": "arena-text-to-image", "benchmarkSnapshotId": "arena-text-to-image-snapshot-2026-10-07", "score": 1383, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-07", "sourceRow": "gpt-image-2 (medium)" }, { "benchmarkId": "arena-image-edit", "benchmarkSnapshotId": "arena-image-edit-snapshot-2026-10-06", "score": 1462, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-06", "sourceRow": "gpt-image-2 (medium)" } ], "officialSourceIds": [ "specialist-openai-gpt-image-2" ] }, { "id": "gemini-3.1-flash-image", "providerId": "google", "name": "Gemini 3.1 Flash Image", "family": "Gemini Image", "releaseDate": "2026-05-28", "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "image", "multimodal" ], "capabilities": [ "image-generation", "image-editing", "image-to-image" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text", "image" ], "strengths": [ "Google documents native text-and-image generation and editing with up to 4K output" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Interleaved text and image generation" ], "pricing": { "inputPerMillionUsd": 0.5, "outputPerMillionUsd": 3, "cachedInputPerMillionUsd": null, "sourceId": "specialist-google-gemini-3-1-flash-image-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.5, "tier": "standard", "sourceId": "specialist-google-gemini-3-1-flash-image-pricing" }, { "unit": "million-output-tokens", "usd": 3, "tier": "standard", "sourceId": "specialist-google-gemini-3-1-flash-image-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-google-gemini-3-1-flash-image-pricing" }, { "unit": "million-output-tokens", "usd": 60, "tier": "standard-image-output", "sourceId": "specialist-google-gemini-3-1-flash-image-pricing" }, { "unit": "image", "usd": 0.045, "tier": "standard-0.5k", "sourceId": "specialist-google-gemini-3-1-flash-image-pricing" }, { "unit": "image", "usd": 0.067, "tier": "standard-1k", "sourceId": "specialist-google-gemini-3-1-flash-image-pricing" }, { "unit": "image", "usd": 0.101, "tier": "standard-2k", "sourceId": "specialist-google-gemini-3-1-flash-image-pricing" }, { "unit": "image", "usd": 0.151, "tier": "standard-4k", "sourceId": "specialist-google-gemini-3-1-flash-image-pricing" } ], "benchmarkResults": [ { "benchmarkId": "arena-text-to-image", "benchmarkSnapshotId": "arena-text-to-image-snapshot-2026-10-07", "score": 1262, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-07", "sourceRow": "gemini-3.1-flash-image (nano-banana-2) [web-search]" }, { "benchmarkId": "arena-image-edit", "benchmarkSnapshotId": "arena-image-edit-snapshot-2026-10-06", "score": 1387, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-06", "sourceRow": "gemini-3.1-flash-image (nano-banana-2) [web-search]" } ], "officialSourceIds": [ "specialist-google-native-image-capabilities", "specialist-google-gemini-3-1-flash-image-pricing", "google-gemini-deprecations" ] }, { "id": "grok-imagine-image", "providerId": "xai", "name": "grok-imagine-image", "family": "Grok Imagine", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "image" ], "capabilities": [ "image-generation", "image-editing", "image-to-image" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "image" ], "strengths": [ "xAI documents text and image inputs for image generation and editing" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Grok Imagine image workflows" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "specialist-xai-imagine-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-xai-imagine-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-xai-imagine-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-xai-imagine-pricing" }, { "unit": "image", "usd": 0.02, "tier": "output-1k", "sourceId": "specialist-xai-imagine-pricing" } ], "benchmarkResults": [ { "benchmarkId": "arena-text-to-image", "benchmarkSnapshotId": "arena-text-to-image-snapshot-2026-10-07", "score": 1170, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-07", "sourceRow": "grok-imagine-image" }, { "benchmarkId": "arena-image-edit", "benchmarkSnapshotId": "arena-image-edit-snapshot-2026-10-06", "score": 1329, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-06", "sourceRow": "grok-imagine-image" } ], "officialSourceIds": [ "specialist-xai-imagine-image", "specialist-xai-imagine-pricing" ] }, { "id": "stable-image-ultra", "providerId": "stability-ai", "name": "Stable Image Ultra", "family": "Stable Image", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "image" ], "capabilities": [ "image-generation" ], "inputModalities": [ "text" ], "outputModalities": [ "image" ], "strengths": [ "Stability documents its flagship image service based on Stable Diffusion 3.5 Large" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Hosted high-detail image generation" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "specialist-stability-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-stability-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-stability-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-stability-pricing" } ], "benchmarkResults": [], "officialSourceIds": [ "specialist-stability-pricing" ] }, { "id": "flux-pro-1.1", "providerId": "black-forest-labs", "name": "FLUX1.1 [pro]", "family": "FLUX.1", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "image" ], "capabilities": [ "image-generation" ], "inputModalities": [ "text" ], "outputModalities": [ "image" ], "strengths": [ "Black Forest Labs documents the flux-pro-1.1 text-to-image endpoint" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Hosted FLUX text-to-image generation" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "specialist-bfl-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-bfl-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-bfl-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-bfl-pricing" }, { "unit": "image", "usd": 0.04, "tier": "standard", "sourceId": "specialist-bfl-pricing" } ], "benchmarkResults": [ { "benchmarkId": "arena-text-to-image", "benchmarkSnapshotId": "arena-text-to-image-snapshot-2026-10-07", "score": 1016, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-07", "sourceRow": "flux-1.1-pro" } ], "officialSourceIds": [ "specialist-bfl-models", "specialist-bfl-pricing" ] }, { "id": "veo-3.1-generate-preview", "providerId": "google", "name": "Veo 3.1 Preview", "family": "Veo", "releaseDate": "2025-10-15", "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image", "video", "audio" ], "taskFamilies": [ "video" ], "capabilities": [ "text-to-video", "image-to-video", "sound-generation" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "video", "audio" ], "strengths": [ "Google lists the exact Veo 3.1 preview identifier for video generation with audio" ], "cautions": [ "Preview models may change and have stricter rate limits" ], "bestFor": [ "Gemini API video generation with audio" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "specialist-google-native-media" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-google-native-media" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-google-native-media" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-google-native-media" }, { "unit": "video-second", "usd": 0.4, "tier": "720p-or-1080p-with-audio", "sourceId": "specialist-google-native-media" } ], "benchmarkResults": [ { "benchmarkId": "arena-text-to-video", "benchmarkSnapshotId": "arena-text-to-video-snapshot-2026-09-22", "score": 1364, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-09-22", "sourceRow": "veo-3.1-audio" }, { "benchmarkId": "arena-image-to-video", "benchmarkSnapshotId": "arena-image-to-video-snapshot-2026-09-22", "score": 1398, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-09-22", "sourceRow": "veo-3.1-audio" } ], "officialSourceIds": [ "specialist-google-native-media", "google-gemini-deprecations" ] }, { "id": "gen4.5", "providerId": "runway", "name": "Gen-4.5", "family": "Runway Gen", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image", "video" ], "taskFamilies": [ "video" ], "capabilities": [ "text-to-video", "image-to-video" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "video" ], "strengths": [ "Runway documents the exact gen4.5 identifier with text or image input and video output" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Runway API video generation" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "specialist-runway-models" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-runway-models" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-runway-models" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-runway-models" } ], "benchmarkResults": [ { "benchmarkId": "arena-text-to-video", "benchmarkSnapshotId": "arena-text-to-video-snapshot-2026-09-22", "score": 1225, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-09-22", "sourceRow": "runway-gen-4.5" } ], "officialSourceIds": [ "specialist-runway-models" ] }, { "id": "grok-imagine-video", "providerId": "xai", "name": "grok-imagine-video", "family": "Grok Imagine", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image", "video" ], "taskFamilies": [ "video" ], "capabilities": [ "text-to-video", "image-to-video", "video-editing" ], "inputModalities": [ "text", "image", "video" ], "outputModalities": [ "video" ], "strengths": [ "xAI documents text, image, and video inputs for video generation and editing" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Grok Imagine video generation and editing" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "specialist-xai-imagine-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-xai-imagine-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-xai-imagine-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-xai-imagine-pricing" }, { "unit": "video-second", "usd": 0.05, "tier": "480p", "sourceId": "specialist-xai-imagine-pricing" } ], "benchmarkResults": [ { "benchmarkId": "arena-text-to-video", "benchmarkSnapshotId": "arena-text-to-video-snapshot-2026-09-22", "score": 1342, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-09-22", "sourceRow": "grok-imagine-video-720p" }, { "benchmarkId": "arena-image-to-video", "benchmarkSnapshotId": "arena-image-to-video-snapshot-2026-09-22", "score": 1415, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-09-22", "sourceRow": "grok-imagine-video-720p" } ], "officialSourceIds": [ "specialist-xai-imagine-video", "specialist-xai-imagine-pricing" ] }, { "id": "gpt-realtime-1.5", "providerId": "openai", "name": "GPT-Realtime-1.5", "family": "GPT Realtime", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": 32000, "maxOutputTokens": 4096, "modalities": [ "text", "image", "audio" ], "taskFamilies": [ "audio", "multimodal" ], "capabilities": [ "audio-understanding", "realtime-voice", "speech-to-text", "text-to-speech" ], "inputModalities": [ "text", "image", "audio" ], "outputModalities": [ "text", "audio" ], "strengths": [ "OpenAI documents realtime text and audio input and output for voice agents" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Realtime voice agents" ], "pricing": { "inputPerMillionUsd": 4, "outputPerMillionUsd": 16, "cachedInputPerMillionUsd": 0.4, "sourceId": "specialist-openai-realtime-1-5" }, "prices": [ { "unit": "million-input-tokens", "usd": 4, "tier": "standard", "sourceId": "specialist-openai-realtime-1-5" }, { "unit": "million-output-tokens", "usd": 16, "tier": "standard", "sourceId": "specialist-openai-realtime-1-5" }, { "unit": "million-input-tokens", "usd": 0.4, "tier": "cached", "sourceId": "specialist-openai-realtime-1-5" } ], "benchmarkResults": [], "officialSourceIds": [ "specialist-openai-realtime-1-5" ] }, { "id": "gpt-audio-1.5", "providerId": "openai", "name": "gpt-audio-1.5", "family": "GPT Audio", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": 128000, "maxOutputTokens": 16384, "modalities": [ "text", "audio" ], "taskFamilies": [ "audio", "multimodal" ], "capabilities": [ "audio-understanding", "speech-to-text", "text-to-speech" ], "inputModalities": [ "text", "audio" ], "outputModalities": [ "text", "audio" ], "strengths": [ "OpenAI documents audio and text inputs and outputs through Chat Completions" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Audio-in and audio-out conversations" ], "pricing": { "inputPerMillionUsd": 2.5, "outputPerMillionUsd": 10, "cachedInputPerMillionUsd": null, "sourceId": "specialist-openai-audio-1-5" }, "prices": [ { "unit": "million-input-tokens", "usd": 2.5, "tier": "standard", "sourceId": "specialist-openai-audio-1-5" }, { "unit": "million-output-tokens", "usd": 10, "tier": "standard", "sourceId": "specialist-openai-audio-1-5" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-openai-audio-1-5" } ], "benchmarkResults": [], "officialSourceIds": [ "specialist-openai-audio-1-5" ] }, { "id": "gpt-4o-transcribe", "providerId": "openai", "name": "GPT-4o Transcribe", "family": "GPT Transcribe", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": 16000, "maxOutputTokens": 2000, "modalities": [ "text", "audio" ], "taskFamilies": [ "audio" ], "capabilities": [ "speech-to-text" ], "inputModalities": [ "text", "audio" ], "outputModalities": [ "text" ], "strengths": [ "OpenAI documents speech-to-text with improved word error rate over Whisper" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Speech transcription" ], "pricing": { "inputPerMillionUsd": 2.5, "outputPerMillionUsd": 10, "cachedInputPerMillionUsd": null, "sourceId": "specialist-openai-transcribe" }, "prices": [ { "unit": "million-input-tokens", "usd": 2.5, "tier": "standard", "sourceId": "specialist-openai-transcribe" }, { "unit": "million-output-tokens", "usd": 10, "tier": "standard", "sourceId": "specialist-openai-transcribe" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-openai-transcribe" } ], "benchmarkResults": [], "officialSourceIds": [ "specialist-openai-transcribe" ] }, { "id": "tts-1", "providerId": "openai", "name": "TTS-1", "family": "OpenAI TTS", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "audio" ], "taskFamilies": [ "audio" ], "capabilities": [ "text-to-speech" ], "inputModalities": [ "text" ], "outputModalities": [ "audio" ], "strengths": [ "OpenAI documents TTS-1 for realtime text-to-speech use cases" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Low-latency text-to-speech" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "specialist-openai-tts-1" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-openai-tts-1" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-openai-tts-1" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-openai-tts-1" } ], "benchmarkResults": [], "officialSourceIds": [ "specialist-openai-tts-1" ] }, { "id": "eleven_v3", "providerId": "elevenlabs", "name": "Eleven v3", "family": "Eleven TTS", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "audio" ], "taskFamilies": [ "audio" ], "capabilities": [ "text-to-speech" ], "inputModalities": [ "text" ], "outputModalities": [ "audio" ], "strengths": [ "ElevenLabs documents expressive speech synthesis across more than 70 languages" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Expressive multilingual speech" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "specialist-elevenlabs-models" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-elevenlabs-models" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-elevenlabs-models" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-elevenlabs-models" } ], "benchmarkResults": [], "officialSourceIds": [ "specialist-elevenlabs-models" ] }, { "id": "eleven_multilingual_v2", "providerId": "elevenlabs", "name": "Eleven Multilingual v2", "family": "Eleven TTS", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "audio" ], "taskFamilies": [ "audio" ], "capabilities": [ "text-to-speech" ], "inputModalities": [ "text" ], "outputModalities": [ "audio" ], "strengths": [ "ElevenLabs documents stable, lifelike multilingual long-form speech" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Long-form multilingual narration" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "specialist-elevenlabs-models" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-elevenlabs-models" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-elevenlabs-models" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-elevenlabs-models" } ], "benchmarkResults": [], "officialSourceIds": [ "specialist-elevenlabs-models" ] }, { "id": "eleven_flash_v2_5", "providerId": "elevenlabs", "name": "Eleven Flash v2.5", "family": "Eleven TTS", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "audio" ], "taskFamilies": [ "audio" ], "capabilities": [ "text-to-speech", "realtime-voice" ], "inputModalities": [ "text" ], "outputModalities": [ "audio" ], "strengths": [ "ElevenLabs documents approximately 75 ms model latency across 32 languages" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Low-latency multilingual speech" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "specialist-elevenlabs-models" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-elevenlabs-models" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-elevenlabs-models" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-elevenlabs-models" } ], "benchmarkResults": [], "officialSourceIds": [ "specialist-elevenlabs-models" ] }, { "id": "scribe_v2", "providerId": "elevenlabs", "name": "Scribe v2", "family": "Scribe", "releaseDate": "2026-01-09", "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "audio", "text" ], "taskFamilies": [ "audio" ], "capabilities": [ "speech-to-text", "diarization", "audio-understanding" ], "inputModalities": [ "audio" ], "outputModalities": [ "text" ], "strengths": [ "ElevenLabs documents transcription in more than 90 languages with diarization" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Batch transcription and diarization" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "specialist-elevenlabs-models" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-elevenlabs-models" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-elevenlabs-models" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-elevenlabs-models" } ], "benchmarkResults": [], "officialSourceIds": [ "specialist-elevenlabs-models", "specialist-elevenlabs-scribe-v2-release" ] }, { "id": "scribe_v2_realtime", "providerId": "elevenlabs", "name": "Scribe v2 Realtime", "family": "Scribe", "releaseDate": "2025-11-11", "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "audio", "text" ], "taskFamilies": [ "audio" ], "capabilities": [ "speech-to-text", "realtime-voice", "audio-understanding" ], "inputModalities": [ "audio" ], "outputModalities": [ "text" ], "strengths": [ "ElevenLabs documents realtime transcription in more than 90 languages" ], "cautions": [ "No exact compatible public benchmark snapshot is recorded" ], "bestFor": [ "Live transcription" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "specialist-elevenlabs-models" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-elevenlabs-models" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "specialist-elevenlabs-models" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "specialist-elevenlabs-models" } ], "benchmarkResults": [], "officialSourceIds": [ "specialist-elevenlabs-models", "specialist-elevenlabs-scribe-realtime-release" ] }, { "id": "gpt-6-astra", "providerId": "openai", "name": "GPT-6 Astra", "family": "GPT-6", "releaseDate": "2026-09-03", "status": "active", "access": "api", "contextWindowTokens": 1050000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "OpenAI positions GPT-6 Astra as its frontier model for reasoning, agentic work, and computer use, with a 1,050,000-token context window and 128,000-token maximum output", "OpenAI reports the strongest published deep-context retention of any frontier model on MRCR v2 (8 needles): 100% up to 512K tokens and 96.3% in the 512K-1M band", "OpenAI reports saturation-level scores on FrontierMath Tier 4 (97.6%) and, under its own provider adapter harness, ARC-AGI-3 (99.9%)" ], "cautions": [ "Long-context billing is a cliff, not a taper: a request over the long-context threshold is rebilled in full at $20 per million input and $75 per million output, so a prompt that crosses it costs roughly double", "GPT-6 Astra is the first publicly released model OpenAI classifies as crossing the Critical cybersecurity threshold under its own Preparedness Framework, which carries additional deployment safeguards", "Vendor-reported benchmark figures come from OpenAI’s own launch material and its own harness; they are not independently reproduced here" ], "bestFor": [ "Frontier reasoning and agentic evaluations where deep-context retrieval accuracy is the binding constraint", "Workloads that can be kept under the long-context threshold, where the effective rate stays at $10 / $50" ], "pricing": { "inputPerMillionUsd": 10, "outputPerMillionUsd": 50, "cachedInputPerMillionUsd": 1, "sourceId": "openai-api-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 10, "tier": "standard", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 50, "tier": "standard", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 1, "tier": "cached", "sourceId": "openai-api-pricing-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 95.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-6-astra_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 93.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-6-astra_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 75.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-6-astra_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1475, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gpt-6-astra-max" } ], "officialSourceIds": [ "openai-gpt-6-astra-model", "openai-api-pricing-2026-09" ] }, { "id": "claude-fable-5-1", "providerId": "anthropic", "name": "Claude Fable 5.1", "family": "Claude Fable", "releaseDate": "2026-09-01", "status": "active", "access": "api", "contextWindowTokens": 1000000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Anthropic’s highest-capability generally available model, for demanding reasoning and long-horizon agentic work, with a 1M-token context window and 128,000-token maximum output", "Cache reads are priced at 0.025x base input ($0.25 per million) rather than the 0.1x every other Claude model uses, which is the single largest cost lever for agentic loops that replay a large prompt", "Anthropic reports 52.6% on Terminal-Bench-Science 0.1 against 24.7% for Fable 5 and 29.0% for Opus 5", "Adaptive thinking is always on and the API effort default is `high`" ], "cautions": [ "Base input and output pricing is unchanged from Fable 5 at $10 / $50; the saving is entirely in cache reads, so a workload with little cache reuse sees no discount", "The 4.7-generation tokenizer produces roughly 30% more tokens for the same text than Claude Sonnet 4.6 and earlier, so per-token rates are not directly comparable across generations", "Anthropic’s launch comparisons are self-reported and were not reproduced independently here" ], "bestFor": [ "Long-horizon agentic coding where a large prompt prefix is replayed across many turns", "Work where Claude Opus 5 at higher effort has already been evaluated and fell short" ], "pricing": { "inputPerMillionUsd": 10, "outputPerMillionUsd": 50, "cachedInputPerMillionUsd": 0.25, "sourceId": "anthropic-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 10, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 50, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.25, "tier": "cached", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 12.5, "tier": "cache-write-5m", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 20, "tier": "cache-write-1h", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 5, "tier": "batch", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 25, "tier": "batch", "sourceId": "anthropic-pricing-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 90.2, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-fable-5-1_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 70.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-fable-5-1_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1501, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "claude-fable-5.1-max" } ], "officialSourceIds": [ "anthropic-models-overview-2026-09", "anthropic-pricing-2026-09" ] }, { "id": "claude-mythos-5-1", "providerId": "anthropic", "name": "Claude Mythos 5.1", "family": "Claude Mythos", "releaseDate": "2026-09-01", "status": "active", "access": "api", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Shipped alongside Claude Fable 5.1 and listed at the same rates: $10 per million input, $50 per million output, and the same 0.025x cache-read multiplier ($0.25 per million)", "Carries the same reduced cache-read pricing as Fable 5.1, which no other Claude model has" ], "cautions": [ "Limited availability: access is restricted to organisations vetted through Anthropic’s Glasswing programme, so it cannot be assumed available for a general evaluation", "Anthropic’s public model comparison table does not list a context window or maximum output for Mythos 5.1, so both are recorded as unknown here rather than inferred from Fable 5.1" ], "bestFor": [ "Organisations already admitted to Anthropic’s limited-availability programme that need frontier capability at Fable 5.1 rates" ], "pricing": { "inputPerMillionUsd": 10, "outputPerMillionUsd": 50, "cachedInputPerMillionUsd": 0.25, "sourceId": "anthropic-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 10, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 50, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.25, "tier": "cached", "sourceId": "anthropic-pricing-2026-09" } ], "benchmarkResults": [], "officialSourceIds": [ "anthropic-pricing-2026-09", "anthropic-glasswing-2026-09" ] }, { "id": "claude-opus-5", "providerId": "anthropic", "name": "Claude Opus 5", "family": "Claude Opus", "releaseDate": "2026-07-24", "status": "active", "access": "api", "contextWindowTokens": 1000000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Anthropic’s recommended default for most workloads, at half Fable 5.1’s token price ($5 / $25) with the same 1M-token context window and 128,000-token maximum output", "Exposes five effort settings (low, medium, high, xhigh, max), which trade capability against token consumption without changing model", "One of only two models with a Fast mode research preview, billed at $10 / $50 across the full context window" ], "cautions": [ "Fast mode pricing applies to the whole request, including above 200K input tokens, so it doubles the effective rate rather than adding a surcharge", "Cache reads are $0.50 per million (the standard 0.1x), five times Fable 5.1’s $0.25, even though Opus 5 is the cheaper model on base tokens. A cache-heavy agent loop can invert the expected cost ordering between the two." ], "bestFor": [ "Complex agentic coding and enterprise work as the general-purpose default before reaching for Fable 5.1", "Latency-sensitive work that can justify Fast mode pricing" ], "pricing": { "inputPerMillionUsd": 5, "outputPerMillionUsd": 25, "cachedInputPerMillionUsd": 0.5, "sourceId": "anthropic-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 5, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 25, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.5, "tier": "cached", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 2.5, "tier": "batch", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 12.5, "tier": "batch", "sourceId": "anthropic-pricing-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 93.9, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-opus-5_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 85.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-opus-5_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 59.9, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-opus-5_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1490, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "claude-opus-5-high" } ], "officialSourceIds": [ "anthropic-models-overview-2026-09", "anthropic-pricing-2026-09" ] }, { "id": "gemini-38-flash", "providerId": "google", "name": "Gemini 3.8 Flash", "family": "Gemini 3.8", "releaseDate": "2026-09-02", "status": "active", "access": "api", "contextWindowTokens": 1048576, "maxOutputTokens": 65536, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "The cheapest frontier-tier option in this dataset by a wide margin: $0.75 per million input and $3.75 per million output, roughly one thirteenth of GPT-6 Astra on both", "Accepts text, image, audio, video and PDF input with a 1,048,576-token context window and 65,536-token maximum output", "Google reports it ahead of Gemini 3.7 Flash on every published benchmark, including DeepSWE v1.1 73.7% vs 65.3% and Terminal-bench 2.1 89.4% vs 85.8%" ], "cautions": [ "The listed price is promotional and ends 31 December 2026; on 1 January 2027 input and output both double, to $1.50 and $7.50, and cached input goes from $0.075 to $0.15", "Output pricing is billed inclusive of thinking tokens, so reasoning-heavy prompts cost more than the output length suggests", "Google’s published comparisons are against its own previous Flash release and were not independently reproduced here" ], "bestFor": [ "High-volume agentic and coding workloads where per-token cost dominates the decision", "Multimodal pipelines that need audio, video and PDF input in one model" ], "pricing": { "inputPerMillionUsd": 0.75, "outputPerMillionUsd": 3.75, "cachedInputPerMillionUsd": 0.075, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.75, "tier": "standard", "effectiveUntil": "2026-12-31", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 1.5, "tier": "standard", "effectiveFrom": "2027-01-01", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 3.75, "tier": "standard", "effectiveUntil": "2026-12-31", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 7.5, "tier": "standard", "effectiveFrom": "2027-01-01", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 0.075, "tier": "cached", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 0.375, "tier": "batch", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 1.875, "tier": "batch", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 1.35, "tier": "priority", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 6.75, "tier": "priority", "sourceId": "google-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 95.4, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.8-flash_high" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 68.4, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.8-flash_high" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 69.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.8-flash_high" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1497, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gemini-3.8-flash-high" } ], "officialSourceIds": [ "google-gemini-38-flash-2026-09", "google-pricing" ] }, { "id": "muse-spark-13", "providerId": "meta", "name": "Muse Spark 1.3", "family": "Muse Spark", "releaseDate": "2026-09-02", "status": "active", "access": "api", "contextWindowTokens": 1048576, "maxOutputTokens": null, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "Meta tunes it for agentic workflows (multi-step tool use, browser control, and long-horizon tasks) with a 1,048,576-token context window", "Meta reports it finishing coding work with roughly 20% fewer tool calls and 25% fewer tokens than Muse Spark 1.2, which compounds with the low token price", "Behavioural changes aimed at real use: it asks clarifying questions on ambiguous prompts and confirms before consequential actions" ], "cautions": [ "Meta’s own documentation states audio understanding is not fully supported and response quality may be degraded on requests containing audio", "The much cheaper contributor endpoint (about $0.10 / $0.20 per million) is priced that way in exchange for Meta training on your traffic. The standard tier is the rate recorded here.", "Meta’s scorecard has it trailing Claude Opus 5 on every broader agent-workflow benchmark it published, including GDPVal-AA v2, despite winning the coding rows" ], "bestFor": [ "Tool-heavy agent loops where fewer calls per task matters as much as the per-token rate", "Coding agents at volume, where its coding scores and low price line up" ], "pricing": { "inputPerMillionUsd": 1.25, "outputPerMillionUsd": 4.25, "cachedInputPerMillionUsd": 0.15, "sourceId": "openrouter-models-api-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 1.25, "tier": "standard", "sourceId": "openrouter-models-api-2026-09" }, { "unit": "million-output-tokens", "usd": 4.25, "tier": "standard", "sourceId": "openrouter-models-api-2026-09" }, { "unit": "million-input-tokens", "usd": 0.15, "tier": "cached", "sourceId": "openrouter-models-api-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 74.4, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "muse-spark-1.3_xhigh" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1494, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "muse-spark-1.3-max" } ], "officialSourceIds": [ "meta-model-api-models-2026-09" ] }, { "id": "qwen38-max", "providerId": "qwen", "name": "Qwen3.8-Max", "family": "Qwen3.8", "releaseDate": "2026-08-03", "status": "active", "access": "api", "contextWindowTokens": 1000000, "maxOutputTokens": 131072, "modalities": [ "text", "image", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image", "video" ], "outputModalities": [ "text" ], "strengths": [ "A 2.4-trillion-parameter mixture-of-experts model activating about 95 billion parameters per query, with a 1M-token context window and reasoning enabled by default", "Output is the cheapest of any frontier-tier text model here at $6 per million, an eighth of GPT-6 Astra and Claude Fable 5.1", "Accepts text, image and video input and returns text", "The 0902 post-training refresh (2 September 2026) targets coding and agent performance specifically" ], "cautions": [ "Two releases share the name: the original 3 August 2026 model and the 0902 refresh. The rates and limits here are read from the 0902 listing", "Reported benchmark placements come from Alibaba and from public arena boards rather than a common pinned snapshot, so they are not comparable with the LiveBench results in this dataset" ], "bestFor": [ "Output-heavy generation workloads where the $6 per million output rate is the deciding factor", "Agentic and coding evaluations that need a 1M-token window at mid-tier cost" ], "pricing": { "inputPerMillionUsd": 2, "outputPerMillionUsd": 6, "cachedInputPerMillionUsd": 0.25, "sourceId": "openrouter-models-api-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 2, "tier": "standard", "sourceId": "openrouter-models-api-2026-09" }, { "unit": "million-output-tokens", "usd": 6, "tier": "standard", "sourceId": "openrouter-models-api-2026-09" }, { "unit": "million-input-tokens", "usd": 0.25, "tier": "cached", "sourceId": "openrouter-models-api-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 92.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "qwen3.8-max_xhigh" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 74.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "qwen3.8-max_xhigh" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 45.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "qwen3.8-max_xhigh" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1483, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "qwen3.8-max" } ], "officialSourceIds": [ "alibaba-model-studio-models-2026-09" ] }, { "id": "kimi-k3", "providerId": "kimi", "name": "Kimi K3", "family": "Kimi K3", "releaseDate": "2026-07-16", "status": "active", "access": "both", "contextWindowTokens": 1048576, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "A 2.8-trillion-parameter mixture-of-experts model with about 104 billion active parameters, a 1M-token context window, and native vision", "Available both as a hosted API and as open weights: the checkpoint shipped 27 July 2026 under the Kimi K3 License as a native MXFP4 quantisation", "Introduces Kimi Delta Attention, a hybrid linear attention scheme, together with attention residuals" ], "cautions": [ "Hosted rates match Claude Sonnet 5 exactly at $3 / $15 while offering a larger context window, so the comparison turns on capability rather than price", "Running the open weights shifts cost to infrastructure, which the token rates recorded here do not represent", "Moonshot’s self-reported results place it ahead of Claude Opus 4.8 and GPT-5.5 but behind Claude Fable 5 and GPT-5.6 Sol; none of this was reproduced independently here" ], "bestFor": [ "Teams that want a frontier-scale model they can also self-host, with one architecture across both paths", "Long-context work at mid-tier hosted pricing" ], "pricing": { "inputPerMillionUsd": 3, "outputPerMillionUsd": 15, "cachedInputPerMillionUsd": 0.3, "sourceId": "openrouter-models-api-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 3, "tier": "standard", "sourceId": "openrouter-models-api-2026-09" }, { "unit": "million-output-tokens", "usd": 15, "tier": "standard", "sourceId": "openrouter-models-api-2026-09" }, { "unit": "million-input-tokens", "usd": 0.3, "tier": "cached", "sourceId": "openrouter-models-api-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 93.1, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "kimi-k3_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 72.2, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "kimi-k3_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 50.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "kimi-k3_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1488, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "kimi-k3-max" } ], "officialSourceIds": [ "kimi-platform-pricing-2026-09" ] }, { "id": "gemini-3-7-flash", "providerId": "google", "name": "Gemini 3.7 Flash", "family": "Gemini 3.7", "releaseDate": "2026-08-13", "status": "active", "access": "api", "apiModelId": "gemini-3.7-flash", "contextWindowTokens": 1048576, "maxOutputTokens": 65536, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning", "coding", "agentic-coding", "instruction-following" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "Google positions Gemini 3.7 Flash for higher-accuracy coding and agent workflows with multimodal inputs and a 1,048,576-token context window", "Google reports strong gains over 3.6 Flash in debugging and issue resolution, more diligent multi-step planning and tool calls, and closer instruction following", "Epoch AI measured 94.8% on GPQA Diamond and 71.6% on FrontierMath Tiers 1-3, ahead of 3.6 Flash on both" ], "cautions": [ "Google publishes the introductory API price through 2026-12-31; standard input and output prices increase on 2027-01-01.", "Not in the pinned LiveBench snapshot; its independent scores here come from Epoch AI's own GPQA Diamond, FrontierMath and SimpleQA Verified runs" ], "bestFor": [ "Production coding and agent workflows that need multimodal context" ], "pricing": { "inputPerMillionUsd": 0.75, "outputPerMillionUsd": 3.75, "cachedInputPerMillionUsd": null, "sourceId": "google-gemini-3-7-flash-release" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.75, "tier": "standard", "effectiveUntil": "2026-12-31", "sourceId": "google-gemini-3-7-flash-release" }, { "unit": "million-output-tokens", "usd": 3.75, "tier": "standard", "effectiveUntil": "2026-12-31", "sourceId": "google-gemini-3-7-flash-release" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-gemini-3-7-flash-release" }, { "unit": "million-input-tokens", "usd": 1.5, "tier": "standard", "effectiveFrom": "2027-01-01", "sourceId": "google-gemini-3-7-flash-release" }, { "unit": "million-output-tokens", "usd": 7.5, "tier": "standard", "effectiveFrom": "2027-01-01", "sourceId": "google-gemini-3-7-flash-release" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 94.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.7-flash_high" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 71.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.7-flash_high" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 69.2, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.7-flash_high" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1486, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gemini-3.7-flash-high" } ], "officialSourceIds": [ "google-gemini-3-7-flash-model", "google-gemini-3-7-flash-release" ] }, { "id": "gemini-3-1-pro-preview", "providerId": "google", "name": "Gemini 3.1 Pro Preview", "family": "Gemini 3.1", "releaseDate": "2026-02-19", "status": "active", "access": "api", "contextWindowTokens": 1048576, "maxOutputTokens": 65536, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning", "coding", "agentic-coding" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "Google documents multimodal inputs, thinking, code execution, and function calling for Gemini 3.1 Pro Preview" ], "cautions": [ "This preview endpoint may change before becoming stable; production use should account for its published prompt-length price tiers" ], "bestFor": [ "Multimodal reasoning and agentic coding evaluations" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 2, "tier": "standard-up-to-200k", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 12, "tier": "standard-up-to-200k", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 0.2, "tier": "cached-up-to-200k", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 4, "tier": "standard-over-200k", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 18, "tier": "standard-over-200k", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 0.4, "tier": "cached-over-200k", "sourceId": "google-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 94.4, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.1-pro-preview_high" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 59.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.1-pro-preview" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 73.5, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.1-pro-preview_high" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1487, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gemini-3.1-pro-preview" } ], "officialSourceIds": [ "google-gemini-3-1-pro-preview-model", "google-gemini-deprecations", "google-pricing" ] }, { "id": "gemini-3-flash-preview", "providerId": "google", "name": "Gemini 3 Flash Preview", "family": "Gemini 3", "releaseDate": "2025-12-17", "status": "active", "access": "api", "contextWindowTokens": 1048576, "maxOutputTokens": 65536, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning", "coding", "agentic-coding" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "Google documents multimodal inputs, thinking, computer use, code execution, and function calling" ], "cautions": [ "This preview endpoint may change before becoming stable and the published price varies by input modality" ], "bestFor": [ "Fast multimodal and agentic coding evaluations" ], "pricing": { "inputPerMillionUsd": 0.5, "outputPerMillionUsd": 3, "cachedInputPerMillionUsd": 0.05, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.5, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 3, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 0.05, "tier": "cached", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 1, "tier": "standard-audio", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 0.1, "tier": "cached-audio", "sourceId": "google-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 89.4, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3-flash-preview_high" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 51.2, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3-flash-preview" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 66.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3-flash-preview_high" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1473, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gemini-3-flash" } ], "officialSourceIds": [ "google-gemini-3-flash-preview-model", "google-gemini-deprecations", "google-pricing" ] }, { "id": "gemini-3-1-flash-lite", "providerId": "google", "name": "Gemini 3.1 Flash-Lite", "family": "Gemini 3.1", "releaseDate": "2026-05-07", "status": "active", "access": "api", "contextWindowTokens": 1048576, "maxOutputTokens": 65536, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning", "translation", "extraction" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "Google documents low-latency multimodal processing for high-volume translation, extraction, and lightweight agentic tasks" ], "cautions": [ "The published price for audio input differs from the listed text, image, and video input price", "Google's earliest published shutdown date is 2027-05-07; migrate to gemini-3.5-flash-lite." ], "bestFor": [ "High-volume multimodal extraction and translation" ], "pricing": { "inputPerMillionUsd": 0.25, "outputPerMillionUsd": 1.5, "cachedInputPerMillionUsd": 0.025, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.25, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 1.5, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 0.025, "tier": "cached", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 0.5, "tier": "standard-audio", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 0.05, "tier": "cached-audio", "sourceId": "google-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 81.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.1-flash-lite_high" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 27.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gemini-3.1-flash-lite_high" } ], "officialSourceIds": [ "google-gemini-3-1-flash-lite-model", "google-gemini-deprecations", "google-pricing" ] }, { "id": "gemini-3-5-live-translate", "providerId": "google", "name": "Gemini 3.5 Live Translate Preview", "family": "Gemini 3.5", "releaseDate": null, "status": "active", "access": "api", "contextWindowTokens": 131072, "maxOutputTokens": 65536, "modalities": [ "text", "audio" ], "taskFamilies": [ "audio" ], "capabilities": [ "realtime-voice", "speech-translation" ], "inputModalities": [ "audio" ], "outputModalities": [ "text", "audio" ], "strengths": [ "Google documents low-latency real-time speech-to-speech translation across 70+ languages" ], "cautions": [ "This preview endpoint supports translation only, rather than tools or instructions, and may change before stable release" ], "bestFor": [ "Realtime spoken-language translation" ], "pricing": { "inputPerMillionUsd": 3.5, "outputPerMillionUsd": 21, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 3.5, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 21, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" }, { "unit": "audio-minute", "usd": 0.0368, "tier": "effective-input-and-output", "sourceId": "google-pricing" } ], "benchmarkResults": [], "officialSourceIds": [ "google-gemini-3-5-live-translate-model", "google-live-api-guide", "google-pricing" ] }, { "id": "gemini-3-1-flash-live", "providerId": "google", "name": "Gemini 3.1 Flash Live Preview", "family": "Gemini 3.1", "releaseDate": "2026-03-26", "status": "active", "access": "api", "contextWindowTokens": 131072, "maxOutputTokens": 65536, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "audio" ], "capabilities": [ "audio-understanding", "realtime-voice", "text-to-speech", "reasoning" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text", "audio" ], "strengths": [ "Google documents a low-latency audio-to-audio model with multimodal awareness for realtime dialogue" ], "cautions": [ "This preview endpoint may change before becoming stable and has input prices that vary by modality" ], "bestFor": [ "Realtime voice applications with multimodal context" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" }, { "unit": "audio-minute", "usd": 0.005, "tier": "input", "sourceId": "google-pricing" }, { "unit": "audio-minute", "usd": 0.018, "tier": "output", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 0.75, "tier": "standard-text", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 4.5, "tier": "standard-text", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 3, "tier": "standard-audio", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 12, "tier": "standard-audio", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 1, "tier": "standard-image-video", "sourceId": "google-pricing" } ], "benchmarkResults": [], "officialSourceIds": [ "google-gemini-3-1-flash-live-model", "google-live-api-guide", "google-gemini-3-5-changelog", "google-pricing" ] }, { "id": "gemini-3-1-flash-tts", "providerId": "google", "name": "Gemini 3.1 Flash TTS Preview", "family": "Gemini 3.1", "releaseDate": "2026-04-15", "status": "active", "access": "api", "contextWindowTokens": 8192, "maxOutputTokens": 16384, "modalities": [ "text", "audio" ], "taskFamilies": [ "audio" ], "capabilities": [ "text-to-speech" ], "inputModalities": [ "text" ], "outputModalities": [ "audio" ], "strengths": [ "Google documents low-latency, controllable speech generation with expressive audio tags" ], "cautions": [ "This preview endpoint may change before becoming stable and has more restrictive rate limits" ], "bestFor": [ "Controllable text-to-speech generation" ], "pricing": { "inputPerMillionUsd": 1, "outputPerMillionUsd": 20, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 1, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 20, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" } ], "benchmarkResults": [], "officialSourceIds": [ "google-gemini-3-1-flash-tts-model", "google-tts-guide", "google-gemini-3-5-changelog", "google-pricing" ] }, { "id": "gemini-3-1-flash-lite-image", "providerId": "google", "name": "Gemini 3.1 Flash Lite Image", "family": "Gemini Image", "releaseDate": "2026-06-30", "status": "active", "access": "api", "contextWindowTokens": 65536, "maxOutputTokens": 4096, "modalities": [ "text", "image" ], "taskFamilies": [ "image" ], "capabilities": [ "image-generation", "image-editing", "image-to-image" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text", "image" ], "strengths": [ "Google documents ultra-low-latency native image generation and editing optimized for 1K output" ], "cautions": [ "Only 1K output is supported; Google documents 2K and 4K as unsupported" ], "bestFor": [ "Low-latency 1K image generation and edits" ], "pricing": { "inputPerMillionUsd": 0.25, "outputPerMillionUsd": 1.5, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.25, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 1.5, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" }, { "unit": "image", "usd": 0.0336, "tier": "standard-1k", "sourceId": "google-pricing" } ], "benchmarkResults": [ { "benchmarkId": "arena-text-to-image", "benchmarkSnapshotId": "arena-text-to-image-snapshot-2026-10-07", "score": 1250, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-07", "sourceRow": "gemini-3.1-flash-lite-image (nano-banana-2-lite)" }, { "benchmarkId": "arena-image-edit", "benchmarkSnapshotId": "arena-image-edit-snapshot-2026-10-06", "score": 1314, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-06", "sourceRow": "gemini-3.1-flash-lite-image (nano-banana-2-lite)" } ], "officialSourceIds": [ "google-gemini-3-1-flash-lite-image-model", "google-gemini-3-5-changelog", "google-pricing" ] }, { "id": "gemini-3-pro-image", "providerId": "google", "name": "Gemini 3 Pro Image", "family": "Gemini Image", "releaseDate": "2026-05-28", "status": "active", "access": "api", "contextWindowTokens": 65536, "maxOutputTokens": 32768, "modalities": [ "text", "image" ], "taskFamilies": [ "image" ], "capabilities": [ "image-generation", "image-editing", "image-to-image" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text", "image" ], "strengths": [ "Google documents native image generation and professional-grade image editing with image and text input" ], "cautions": [ "Pricing differs by standard, batch, flex, and priority consumption" ], "bestFor": [ "Professional image generation and editing" ], "pricing": { "inputPerMillionUsd": 2, "outputPerMillionUsd": 12, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 2, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 12, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" }, { "unit": "image", "usd": 0.134, "tier": "standard-1k-or-2k", "sourceId": "google-pricing" }, { "unit": "image", "usd": 0.24, "tier": "standard-4k", "sourceId": "google-pricing" } ], "benchmarkResults": [ { "benchmarkId": "arena-text-to-image", "benchmarkSnapshotId": "arena-text-to-image-snapshot-2026-10-07", "score": 1248, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-07", "sourceRow": "gemini-3-pro-image-2k (nano-banana-pro)" }, { "benchmarkId": "arena-image-edit", "benchmarkSnapshotId": "arena-image-edit-snapshot-2026-10-06", "score": 1390, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-06", "sourceRow": "gemini-3-pro-image-2k (nano-banana-pro)" } ], "officialSourceIds": [ "google-gemini-3-pro-image-model", "google-gemini-deprecations", "google-gemini-3-5-changelog", "google-pricing" ] }, { "id": "gemini-omni-flash", "providerId": "google", "name": "Gemini Omni Flash Preview", "family": "Gemini Omni", "releaseDate": "2026-06-30", "status": "active", "access": "api", "contextWindowTokens": 1048576, "maxOutputTokens": null, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "video" ], "capabilities": [ "text-to-video", "image-to-video", "video-editing", "sound-generation" ], "inputModalities": [ "text", "image", "video" ], "outputModalities": [ "video", "audio" ], "strengths": [ "Google documents conversational video generation and editing from text and images through the Interactions API" ], "cautions": [ "This preview endpoint may change before becoming stable; it generates 3 to 10 second 720p video" ], "bestFor": [ "Conversational video generation and editing" ], "pricing": { "inputPerMillionUsd": 1.5, "outputPerMillionUsd": 9, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 1.5, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 9, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" }, { "unit": "video-second", "usd": 0.1, "tier": "standard-720p", "sourceId": "google-pricing" } ], "benchmarkResults": [ { "benchmarkId": "arena-text-to-video", "benchmarkSnapshotId": "arena-text-to-video-snapshot-2026-09-22", "score": 1513, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-09-22", "sourceRow": "gemini-omni-flash" }, { "benchmarkId": "arena-image-to-video", "benchmarkSnapshotId": "arena-image-to-video-snapshot-2026-09-22", "score": 1465, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-09-22", "sourceRow": "gemini-omni-flash" } ], "officialSourceIds": [ "google-gemini-omni-flash-model", "google-omni-flash-guide", "google-gemini-3-5-changelog", "google-pricing" ] }, { "id": "claude-opus-55", "providerId": "anthropic", "name": "Claude Opus 5.5", "family": "Claude Opus", "releaseDate": "2026-09-22", "status": "active", "access": "api", "contextWindowTokens": 1000000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Anthropic's recommended default for most workloads, taking that role from Claude Opus 5, at $4 per million input and $20 per million output", "1M-token context window with a 128,000-token maximum output, and adaptive thinking that is always on", "Cache reads are 0.05x base input ($0.20 per million) rather than the usual 0.1x, which matters most in agent loops that replay a large prefix" ], "cautions": [ "The API effort default is medium here, not high as on Claude Fable 5.1 and Claude Sonnet 5.5, so a like-for-like comparison has to set effort explicitly", "The 4.7-generation tokenizer produces roughly 30% more tokens for the same text than Claude Sonnet 4.6 and earlier, so per-token rates do not compare directly across generations" ], "bestFor": [ "Long-running agentic coding and knowledge work as the general-purpose default", "Cache-heavy agent loops where the 0.05x read multiplier compounds" ], "pricing": { "inputPerMillionUsd": 4, "outputPerMillionUsd": 20, "cachedInputPerMillionUsd": 0.2, "sourceId": "anthropic-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 4, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 20, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.2, "tier": "cached", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 5, "tier": "cache-write-5m", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 8, "tier": "cache-write-1h", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 2, "tier": "batch", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 10, "tier": "batch", "sourceId": "anthropic-pricing-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 90.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-opus-5-5_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 91.2, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-opus-5-5_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 72.2, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-opus-5-5_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1507, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "claude-opus-5.5-high" } ], "officialSourceIds": [ "anthropic-models-overview-2026-09", "anthropic-pricing-2026-09" ] }, { "id": "claude-sonnet-55", "providerId": "anthropic", "name": "Claude Sonnet 5.5", "family": "Claude Sonnet", "releaseDate": "2026-09-28", "status": "active", "access": "api", "contextWindowTokens": 1000000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Anthropic's speed-and-intelligence tier at $2 per million input and $10 per million output, with the same 1M-token window and 128,000-token output cap as Opus 5.5", "Cache reads at 0.05x base input ($0.10 per million), half the usual multiplier", "Keeps the high effort default, unlike Claude Opus 5.5" ], "cautions": [ "Same list price as Claude Sonnet 5, so the gain is capability and cache economics rather than a cheaper rate card", "Anthropic positions Opus 5.5 rather than this as the default, so reach for Sonnet 5.5 when latency matters more than peak capability" ], "bestFor": [ "Production workloads that need frontier-adjacent quality at a mid-tier rate", "Latency-sensitive agent steps inside a larger pipeline" ], "pricing": { "inputPerMillionUsd": 2, "outputPerMillionUsd": 10, "cachedInputPerMillionUsd": 0.1, "sourceId": "anthropic-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 2, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 10, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.1, "tier": "cached", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 2.5, "tier": "cache-write-5m", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 4, "tier": "cache-write-1h", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 1, "tier": "batch", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 5, "tier": "batch", "sourceId": "anthropic-pricing-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 95.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-sonnet-5-5_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 88.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-sonnet-5-5_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 46.5, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-sonnet-5-5_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1476, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "claude-sonnet-5.5-xhigh" } ], "officialSourceIds": [ "anthropic-models-overview-2026-09", "anthropic-pricing-2026-09" ] }, { "id": "claude-haiku-55", "providerId": "anthropic", "name": "Claude Haiku 5.5", "family": "Claude Haiku", "releaseDate": "2026-10-07", "status": "active", "access": "api", "contextWindowTokens": 1000000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "The cheapest model in this catalogue at $0.10 per million input and $0.50 per million output for prompts up to 100,000 tokens", "Carries the full 1M-token context window and a 128,000-token output cap, which earlier Haiku releases did not", "Cache reads at $0.01 per million make high-volume classification and routing nearly free on the input side" ], "cautions": [ "Priced by prompt length, and it is a cliff rather than a taper: a request whose prompt passes 100,000 tokens is billed at $0.50 input and $2.50 output for the whole request, five times the short rate", "Cache reads and cache writes both count toward that 100,000-token measurement, so a large cached prefix can tip a request into the higher band on its own", "It is the only current Claude model whose 1M window is not priced uniformly; every other 4.6-and-later model bills a 900k request at the same rate as a 9k one" ], "bestFor": [ "High-volume classification, extraction and routing under the 100,000-token line", "Cost-sensitive fan-out steps where a frontier model is overkill" ], "pricing": { "inputPerMillionUsd": 0.1, "outputPerMillionUsd": 0.5, "cachedInputPerMillionUsd": 0.01, "sourceId": "anthropic-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.1, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 0.5, "tier": "standard", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.01, "tier": "cached", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.125, "tier": "cache-write-5m", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.2, "tier": "cache-write-1h", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.5, "tier": "long-prompt-over-100k", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 2.5, "tier": "long-prompt-over-100k", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.05, "tier": "batch", "sourceId": "anthropic-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 0.25, "tier": "batch", "sourceId": "anthropic-pricing-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 89.6, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-haiku-5-5_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 75.1, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-haiku-5-5_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 23.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "claude-haiku-5-5_max" } ], "officialSourceIds": [ "anthropic-models-overview-2026-09", "anthropic-pricing-2026-09" ] }, { "id": "gpt-6-sol", "providerId": "openai", "name": "GPT-6 Sol", "family": "GPT-6", "releaseDate": "2026-09-22", "status": "active", "access": "api", "contextWindowTokens": 1050000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "The mid tier of the GPT-6 family at $2 per million input and $10 per million output, a fifth of GPT-6 Astra on both", "Same 1,050,000-token context window and 128,000-token output cap as GPT-6 Astra" ], "cautions": [ "Prompts over 272,000 input tokens are rebilled for the whole request at $4 input and $15 output, so crossing the line roughly doubles input cost", "GPT-6.1 Sol lists the same base rates with half the cached-input price, so there is little reason to pick this one for a cache-heavy workload" ], "bestFor": [ "General production work that does not need Astra's capability or price", "Workloads kept under the 272,000-token threshold" ], "pricing": { "inputPerMillionUsd": 2, "outputPerMillionUsd": 10, "cachedInputPerMillionUsd": 0.2, "sourceId": "openai-api-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 2, "tier": "standard", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 10, "tier": "standard", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.2, "tier": "cached", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 4, "tier": "long-context-over-272k", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 15, "tier": "long-context-over-272k", "sourceId": "openai-api-pricing-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 94.3, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-6-sol_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 89.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-6-sol_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 60.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-6-sol_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1456, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gpt-6-sol-max" } ], "officialSourceIds": [ "openai-api-pricing-2026-09", "openai-gpt-6-astra-model" ] }, { "id": "gpt-61-sol", "providerId": "openai", "name": "GPT-6.1 Sol", "family": "GPT-6.1", "releaseDate": "2026-09-29", "status": "active", "access": "api", "contextWindowTokens": 1050000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "Matches GPT-6 Sol's $2 / $10 base rates while halving cached input to $0.10 per million, the cheapest cache read of any OpenAI model here", "1,050,000-token context window with a 128,000-token maximum output" ], "cautions": [ "The same 272,000-token long-context cliff applies, rebilling the full request at $4 input and $15 output", "Released a week after GPT-6 Sol at identical base pricing, so the two are easy to confuse on a rate card; the cached rate is the difference" ], "bestFor": [ "Agent loops that replay a large cached prefix", "A drop-in alternative to GPT-6 Sol at the same base cost" ], "pricing": { "inputPerMillionUsd": 2, "outputPerMillionUsd": 10, "cachedInputPerMillionUsd": 0.1, "sourceId": "openai-api-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 2, "tier": "standard", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 10, "tier": "standard", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.1, "tier": "cached", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 4, "tier": "long-context-over-272k", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 15, "tier": "long-context-over-272k", "sourceId": "openai-api-pricing-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 95.4, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-6.1-sol_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 93.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-6.1-sol_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 73.9, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-6.1-sol_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1484, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gpt-6.1-sol-max" } ], "officialSourceIds": [ "openai-api-pricing-2026-09", "openai-gpt-6-astra-model" ] }, { "id": "gpt-6-luna", "providerId": "openai", "name": "GPT-6 Luna", "family": "GPT-6", "releaseDate": "2026-09-22", "status": "active", "access": "api", "contextWindowTokens": 1050000, "maxOutputTokens": 128000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "OpenAI's budget tier at $0.10 per million input and $0.50 per million output, matching Claude Haiku 5.5 on both", "Keeps the full 1,050,000-token context window rather than cutting it down for the cheap tier", "The long-context surcharge starts at 272,000 tokens rather than Claude Haiku 5.5's 100,000, which is the more forgiving threshold of the two budget options" ], "cautions": [ "Prompts over 272,000 tokens are rebilled at $0.20 input and $0.75 output for the whole request", "A twentieth of GPT-6 Sol's price reflects a capability gap, not a discount on the same model" ], "bestFor": [ "High-volume summarisation, extraction and routing", "Long-context budget work, where the 272k threshold is more forgiving than Haiku 5.5's 100k" ], "pricing": { "inputPerMillionUsd": 0.1, "outputPerMillionUsd": 0.5, "cachedInputPerMillionUsd": 0.01, "sourceId": "openai-api-pricing-2026-09" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.1, "tier": "standard", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 0.5, "tier": "standard", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.01, "tier": "cached", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-input-tokens", "usd": 0.2, "tier": "long-context-over-272k", "sourceId": "openai-api-pricing-2026-09" }, { "unit": "million-output-tokens", "usd": 0.75, "tier": "long-context-over-272k", "sourceId": "openai-api-pricing-2026-09" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 90.5, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-6-luna_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 78.9, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-6-luna_max" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 41.4, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "gpt-6-luna_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1443, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "gpt-6-luna-max" } ], "officialSourceIds": [ "openai-api-pricing-2026-09", "openai-gpt-6-astra-model" ] }, { "id": "grok-47", "providerId": "xai", "name": "Grok 4.7", "family": "Grok 4", "releaseDate": "2026-09-21", "status": "active", "access": "api", "contextWindowTokens": 500000, "maxOutputTokens": null, "modalities": [ "text" ], "taskFamilies": [ "text-nlp" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text" ], "outputModalities": [ "text" ], "strengths": [ "Holds Grok 4.6's $2 / $6 rates and 500,000-token context window", "Output at $6 per million undercuts every frontier-tier model here except the explicit budget tiers" ], "cautions": [ "The long-prompt threshold is 200,000 tokens, the lowest of any model here, and reaching it doubles input and output for the entire request", "xAI's model page does not state a maximum output for Grok 4.7, so it is recorded as unknown rather than inferred", "That page does not explicitly list image input for this model either, so it is catalogued as text-only" ], "bestFor": [ "Output-heavy generation under the 200,000-token prompt line", "A cheaper alternative to the frontier tiers where 500k of context is enough" ], "pricing": { "inputPerMillionUsd": 2, "outputPerMillionUsd": 6, "cachedInputPerMillionUsd": 0.5, "sourceId": "xai-models-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 2, "tier": "standard", "sourceId": "xai-models-pricing" }, { "unit": "million-output-tokens", "usd": 6, "tier": "standard", "sourceId": "xai-models-pricing" }, { "unit": "million-input-tokens", "usd": 0.5, "tier": "cached", "sourceId": "xai-models-pricing" }, { "unit": "million-input-tokens", "usd": 4, "tier": "long-prompt-over-200k", "sourceId": "xai-models-pricing" }, { "unit": "million-output-tokens", "usd": 12, "tier": "long-prompt-over-200k", "sourceId": "xai-models-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 92.7, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.7_xhigh" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 53, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.7_xhigh" }, { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 56, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "grok-4.7_xhigh" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1443, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "grok-4.7-xhigh" } ], "officialSourceIds": [ "xai-models-pricing" ] }, { "id": "gemini-38-flash-tts", "providerId": "google", "name": "Gemini 3.8 Flash TTS", "family": "Gemini 3.8", "releaseDate": "2026-09-22", "status": "active", "access": "api", "apiModelId": "gemini-3.8-flash-tts", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "audio" ], "taskFamilies": [ "audio" ], "capabilities": [ "text-to-speech" ], "inputModalities": [ "text" ], "outputModalities": [ "audio" ], "strengths": [ "Google's flagship expressive text-to-speech model, built for studio-grade voice fidelity, with custom voices and multi-speaker scene dialogue from text prompts", "Priced per token rather than per character: $0.50 per million text input tokens and $9.00 per million audio output tokens", "Shares the Gemini 3.8 Flash generation, so a text pipeline and its narration run on the same family" ], "cautions": [ "Google doubles the rates on 2027-01-01, to $1.00 input and $18 output per 1M tokens (about $0.027 per minute)", "Audio output is billed in audio tokens, which do not convert one-for-one to text tokens, so a text-token budget does not predict this cost", "Google's pricing page lists no context window or output cap for this model, so both are recorded as unknown", "No independent speech-quality benchmark is recorded here" ], "bestFor": [ "Narration, audiobooks and product audio where voice quality matters more than cost", "Narration and voice output attached to an existing Gemini text pipeline", "Audio generation budgeted before the 2027 price step" ], "pricing": { "inputPerMillionUsd": 0.5, "outputPerMillionUsd": 9, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.5, "tier": "standard", "effectiveUntil": "2026-12-31", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 9, "tier": "standard", "effectiveUntil": "2026-12-31", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 1, "tier": "standard", "effectiveFrom": "2027-01-01", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 18, "tier": "standard", "effectiveFrom": "2027-01-01", "sourceId": "google-pricing" }, { "unit": "audio-minute", "usd": 0.0135, "tier": "standard", "effectiveUntil": "2026-12-31", "sourceId": "google-pricing" }, { "unit": "audio-minute", "usd": 0.027, "tier": "standard", "effectiveFrom": "2027-01-01", "sourceId": "google-pricing" } ], "benchmarkResults": [], "officialSourceIds": [ "google-gemini-3-8-tts-release", "google-gemini-3-5-changelog", "google-pricing", "google-gemini-38-flash-2026-09" ] }, { "id": "deepseek-v41-flash", "providerId": "deepseek", "name": "DeepSeek-V4.1-Flash", "family": "DeepSeek-V4.1", "releaseDate": "2026-09-10", "status": "active", "access": "both", "apiModelId": "deepseek-flash", "contextWindowTokens": 1000000, "maxOutputTokens": 384000, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "A 552B-parameter mixture-of-experts model that activates 8B parameters for input and 16B for output, with native visual understanding and a 1M-token context window", "Released as MIT-licensed open weights on Hugging Face as well as on the DeepSeek API, so the same model can be hosted or self-run", "DeepSeek says it beats its own former flagship, DeepSeek-V4-Pro, and halves KV-cache memory needs; Epoch AI measured 89.8% on GPQA Diamond" ], "cautions": [ "The recorded API rates are the peak-hour prices; DeepSeek charges half outside 01:00-04:00 and 06:00-10:00 UTC on weekdays", "Running the open weights moves cost to your own infrastructure, which the token rates here do not represent" ], "bestFor": [ "High-volume agent and multimodal workloads where per-token cost dominates", "Teams that want a frontier-adjacent model they can also self-host under MIT terms" ], "pricing": { "inputPerMillionUsd": 0.3, "outputPerMillionUsd": 1.2, "cachedInputPerMillionUsd": 0.006, "sourceId": "deepseek-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.3, "tier": "standard", "sourceId": "deepseek-pricing" }, { "unit": "million-output-tokens", "usd": 1.2, "tier": "standard", "sourceId": "deepseek-pricing" }, { "unit": "million-input-tokens", "usd": 0.006, "tier": "cached", "sourceId": "deepseek-pricing" }, { "unit": "million-input-tokens", "usd": 0.15, "tier": "off-peak", "sourceId": "deepseek-pricing" }, { "unit": "million-output-tokens", "usd": 0.6, "tier": "off-peak", "sourceId": "deepseek-pricing" }, { "unit": "million-input-tokens", "usd": 0.003, "tier": "off-peak-cached", "sourceId": "deepseek-pricing" } ], "benchmarkResults": [ { "benchmarkId": "epoch-gpqa-diamond", "benchmarkSnapshotId": "epoch-gpqa-diamond-snapshot-2026-10-12", "score": 89.8, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "deepseek-v4.1-flash_max" }, { "benchmarkId": "epoch-frontiermath-tiers-1-3-v2", "benchmarkSnapshotId": "epoch-frontiermath-tiers-1-3-v2-snapshot-2026-10-12", "score": 67.4, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "deepseek-v4.1-flash_max" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1475, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "deepseek-v4.1-flash-max" } ], "officialSourceIds": [ "deepseek-v4-1-flash-release", "deepseek-v4-1-flash-weights", "deepseek-pricing" ] }, { "id": "mistral-large-4", "providerId": "mistral", "name": "Mistral Large 4", "family": "Mistral Large 4", "releaseDate": "2026-10-06", "status": "active", "access": "api", "apiModelId": "mistral-large-4-0", "contextWindowTokens": 1000000, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "text-nlp", "multimodal" ], "capabilities": [ "reasoning" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "strengths": [ "A natively multimodal mixture-of-experts model with 1.05 trillion total and 52 billion active parameters, a 1.6B vision encoder and a 1M-token context window", "Mistral reports fluency in more than 160 languages, including every official EU language, and serves the preview from European data centres", "Mistral says open weights follow by the end of October 2026, which would make it one of the largest open-weight models available" ], "cautions": [ "This is a public preview: Mistral says the reinforcement learning run is still in flight and the model will keep changing", "Weights and their licence were not published at the time of this update, so it is recorded as API-only", "Mistral offered a temporary 50% launch discount; the recorded rates are the list prices", "Epoch AI measured 20.0% on SimpleQA Verified, low for its size, so check factual recall on your own questions" ], "bestFor": [ "European data-residency workloads that want a frontier-scale multimodal model", "Teams planning to self-host once the open weights ship" ], "pricing": { "inputPerMillionUsd": 1.36, "outputPerMillionUsd": 4.18, "cachedInputPerMillionUsd": 0.14, "sourceId": "mistral-large-4-model" }, "prices": [ { "unit": "million-input-tokens", "usd": 1.36, "tier": "standard", "sourceId": "mistral-large-4-model" }, { "unit": "million-output-tokens", "usd": 4.18, "tier": "standard", "sourceId": "mistral-large-4-model" }, { "unit": "million-input-tokens", "usd": 0.14, "tier": "cached", "sourceId": "mistral-large-4-model" } ], "benchmarkResults": [ { "benchmarkId": "epoch-simpleqa-verified", "benchmarkSnapshotId": "epoch-simpleqa-verified-snapshot-2026-10-12", "score": 20, "sourceId": "epoch-benchmark-data-2026-10-12", "measuredAt": "2026-10-12", "sourceRow": "mistral-large-4-0_high" }, { "benchmarkId": "arena-text", "benchmarkSnapshotId": "arena-text-snapshot-2026-10-08", "score": 1429, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-08", "sourceRow": "mistral-large-4" } ], "officialSourceIds": [ "mistral-large-4-release", "mistral-large-4-model" ] }, { "id": "qwen38-omni-flash", "providerId": "qwen", "name": "Qwen3.8-Omni-Flash", "family": "Qwen3.8", "releaseDate": "2026-09-20", "status": "active", "access": "api", "apiModelId": "qwen3.8-omni-flash", "contextWindowTokens": 1000000, "maxOutputTokens": 131072, "modalities": [ "text", "image", "audio", "video" ], "taskFamilies": [ "multimodal", "audio" ], "capabilities": [ "reasoning", "audio-understanding", "video-understanding", "image-understanding" ], "inputModalities": [ "text", "image", "audio", "video" ], "outputModalities": [ "text" ], "strengths": [ "An omnimodal model that reads text, images, audio and video in one 1M-token context and answers in text, with thinking, tool calling and web search", "Alibaba positions it for agents that plan, call tools and deliver finished work from audio-visual material, including up to an hour of combined audio and video", "Very low international API rates of $0.15 input and $0.47 output per 1M tokens" ], "cautions": [ "Output is text only; for spoken replies Alibaba lists a separate realtime variant", "Alibaba's comparisons with Gemini 3.8 Flash are self-reported and have not been reproduced independently here", "The release date is the Alibaba Cloud announcement date, 2026-09-20" ], "bestFor": [ "Long audio and video understanding at low cost", "Agent pipelines that must reason over meetings, calls or recorded media" ], "pricing": { "inputPerMillionUsd": 0.15, "outputPerMillionUsd": 0.47, "cachedInputPerMillionUsd": 0.016, "sourceId": "alibaba-model-studio-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.15, "tier": "standard", "sourceId": "alibaba-model-studio-pricing" }, { "unit": "million-output-tokens", "usd": 0.47, "tier": "standard", "sourceId": "alibaba-model-studio-pricing" }, { "unit": "million-input-tokens", "usd": 0.016, "tier": "cached", "sourceId": "alibaba-model-studio-pricing" } ], "benchmarkResults": [], "officialSourceIds": [ "qwen-3-8-omni-flash-model", "qwen-3-8-omni-flash-blog", "alibaba-model-studio-pricing" ] }, { "id": "gemini-nano-banana-2-1", "providerId": "google", "name": "Nano Banana 2.1", "family": "Nano Banana", "releaseDate": "2026-10-06", "status": "active", "access": "api", "apiModelId": "gemini-nano-banana-2.1", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image", "video" ], "taskFamilies": [ "image", "multimodal" ], "capabilities": [ "image-generation", "image-editing", "image-to-image" ], "inputModalities": [ "text", "image", "video" ], "outputModalities": [ "image", "text" ], "strengths": [ "An update to Nano Banana 2 (Gemini 3.1 Flash Image) with better visual quality, prompt adherence, multi-turn character consistency and text rendering at Flash-tier speed", "Generates and edits at 1K, 2K and 4K, accepts text, image, video and PDF input, and fuses up to 14 reference images", "At $0.0336 per 1K image it costs about half of its predecessor's $0.067" ], "cautions": [ "Google prices output by image tokens; the per-image figures are Google's own equivalents and batch is half price", "No independent image-quality benchmark is recorded here" ], "bestFor": [ "High-volume product, marketing and social imagery where cost per image matters", "Iterative editing that must keep characters and objects consistent across turns" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 30, "tier": "standard-image-output", "sourceId": "google-pricing" }, { "unit": "image", "usd": 0.0336, "tier": "standard-1k", "sourceId": "google-pricing" }, { "unit": "image", "usd": 0.0504, "tier": "standard-2k", "sourceId": "google-pricing" }, { "unit": "image", "usd": 0.113, "tier": "standard-4k", "sourceId": "google-pricing" } ], "benchmarkResults": [ { "benchmarkId": "arena-text-to-image", "benchmarkSnapshotId": "arena-text-to-image-snapshot-2026-10-07", "score": 1328, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-07", "sourceRow": "gemini-nano-banana-2.1" }, { "benchmarkId": "arena-image-edit", "benchmarkSnapshotId": "arena-image-edit-snapshot-2026-10-06", "score": 1428, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-06", "sourceRow": "gemini-nano-banana-2.1" } ], "officialSourceIds": [ "google-nano-banana-2-1-model", "google-gemini-3-5-changelog", "google-pricing" ] }, { "id": "gemini-3-8-live", "providerId": "google", "name": "Gemini 3.8 Live", "family": "Gemini 3.8", "releaseDate": "2026-09-15", "status": "active", "access": "api", "apiModelId": "gemini-3.8-live", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "audio" ], "taskFamilies": [ "audio" ], "capabilities": [ "audio-understanding", "realtime-voice", "text-to-speech", "reasoning" ], "inputModalities": [ "text", "audio" ], "outputModalities": [ "text", "audio" ], "strengths": [ "Google's default speech-to-speech model for low-latency voice agents in the Gemini API and Google AI Studio", "A sibling, Gemini 3.8 Live Extended Thinking, trades latency for multi-step reasoning at the same published rates" ], "cautions": [ "Audio is billed per token; the per-minute figures are Google's equivalents ($0.005 input, $0.018 output)", "No independent voice benchmark is recorded here" ], "bestFor": [ "Customer-facing voice agents that need natural turn-taking" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" }, { "unit": "audio-minute", "usd": 0.005, "tier": "input", "sourceId": "google-pricing" }, { "unit": "audio-minute", "usd": 0.018, "tier": "output", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 0.75, "tier": "standard-text", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 4.5, "tier": "standard-text", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 3, "tier": "standard-audio", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 12, "tier": "standard-audio", "sourceId": "google-pricing" } ], "benchmarkResults": [], "officialSourceIds": [ "google-gemini-3-8-live-release", "google-gemini-3-5-changelog", "google-pricing" ] }, { "id": "gemini-3-8-flash-lite-tts", "providerId": "google", "name": "Gemini 3.8 Flash-Lite TTS", "family": "Gemini 3.8", "releaseDate": "2026-09-22", "status": "active", "access": "api", "apiModelId": "gemini-3.8-flash-lite-tts", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "audio" ], "taskFamilies": [ "audio" ], "capabilities": [ "text-to-speech" ], "inputModalities": [ "text" ], "outputModalities": [ "audio" ], "strengths": [ "A faster, cheaper text-to-speech model that replaces Gemini 3.1 Flash TTS Preview, with the same custom-voice and scene-dialogue controls", "Introductory output price of $6 per 1M audio tokens, about $0.009 per minute of speech, through 2026-12-31" ], "cautions": [ "Google doubles the rates on 2027-01-01, to $1.00 input and $12 output per 1M tokens (about $0.018 per minute)", "No independent speech-quality benchmark is recorded here" ], "bestFor": [ "High-volume speech generation such as notifications, IVR prompts and accessibility audio" ], "pricing": { "inputPerMillionUsd": 0.5, "outputPerMillionUsd": 6, "cachedInputPerMillionUsd": null, "sourceId": "google-pricing" }, "prices": [ { "unit": "million-input-tokens", "usd": 0.5, "tier": "standard", "effectiveUntil": "2026-12-31", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 6, "tier": "standard", "effectiveUntil": "2026-12-31", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "google-pricing" }, { "unit": "million-input-tokens", "usd": 1, "tier": "standard", "effectiveFrom": "2027-01-01", "sourceId": "google-pricing" }, { "unit": "million-output-tokens", "usd": 12, "tier": "standard", "effectiveFrom": "2027-01-01", "sourceId": "google-pricing" }, { "unit": "audio-minute", "usd": 0.009, "tier": "standard", "effectiveUntil": "2026-12-31", "sourceId": "google-pricing" }, { "unit": "audio-minute", "usd": 0.018, "tier": "standard", "effectiveFrom": "2027-01-01", "sourceId": "google-pricing" } ], "benchmarkResults": [], "officialSourceIds": [ "google-gemini-3-8-tts-release", "google-gemini-3-5-changelog", "google-pricing" ] }, { "id": "gpt-live-1", "providerId": "openai", "name": "GPT-Live 1", "family": "GPT-Live", "releaseDate": "2026-09-10", "status": "active", "access": "api", "apiModelId": "gpt-live-1", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "audio" ], "taskFamilies": [ "audio" ], "capabilities": [ "audio-understanding", "realtime-voice" ], "inputModalities": [ "text", "audio" ], "outputModalities": [ "text", "audio" ], "strengths": [ "The full-duplex voice model from ChatGPT: it can listen and speak at the same time instead of strictly taking turns", "Hands reasoning and tool use to a backend model, so a voice front end can be paired with GPT-6-class reasoning", "Flat $0.05 per session minute, billed per second" ], "cautions": [ "Backend reasoning and tool calls are billed separately at the configured model's rates, so real cost per minute is higher than the session rate", "OpenAI's task-success comparisons are self-reported and not reproduced independently here" ], "bestFor": [ "Phone and in-app voice agents that need natural interruptions" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "openai-gpt-live-1-model" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "openai-gpt-live-1-model" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "openai-gpt-live-1-model" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "openai-gpt-live-1-model" }, { "unit": "realtime-minute", "usd": 0.05, "tier": "standard", "sourceId": "openai-gpt-live-1-model" } ], "benchmarkResults": [], "officialSourceIds": [ "openai-gpt-live-1-model", "openai-api-changelog" ] }, { "id": "gpt-image-2-5-sunburst", "providerId": "openai", "name": "GPT Image 2.5 Sunburst", "family": "GPT Image", "releaseDate": "2026-09-08", "status": "active", "access": "api", "apiModelId": "gpt-image-2.5-sunburst", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "image" ], "capabilities": [ "image-generation", "image-editing", "image-to-image" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "image" ], "strengths": [ "OpenAI's most capable image generation and editing model, with six quality settings from low to max", "Token rates match GPT Image 2: $5 per 1M text input tokens, $8 per 1M image input tokens and $30 per 1M image output tokens, half that in batch" ], "cautions": [ "OpenAI publishes no per-image price, and its GPT Image 2 calculator does not estimate 2.5 token use, so cost per image has to be measured", "No independent image-quality benchmark is recorded here" ], "bestFor": [ "Hero imagery, detailed edits and text-heavy graphics where quality matters most" ], "pricing": { "inputPerMillionUsd": 5, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": 1.25, "sourceId": "openai-gpt-image-2-5-sunburst-model" }, "prices": [ { "unit": "million-input-tokens", "usd": 5, "tier": "standard", "sourceId": "openai-gpt-image-2-5-sunburst-model" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "openai-gpt-image-2-5-sunburst-model" }, { "unit": "million-input-tokens", "usd": 1.25, "tier": "cached", "sourceId": "openai-gpt-image-2-5-sunburst-model" }, { "unit": "million-input-tokens", "usd": 8, "tier": "standard-image-input", "sourceId": "openai-gpt-image-2-5-sunburst-model" }, { "unit": "million-output-tokens", "usd": 30, "tier": "standard-image-output", "sourceId": "openai-gpt-image-2-5-sunburst-model" }, { "unit": "million-output-tokens", "usd": 15, "tier": "batch-image-output", "sourceId": "openai-gpt-image-2-5-sunburst-model" } ], "benchmarkResults": [ { "benchmarkId": "arena-text-to-image", "benchmarkSnapshotId": "arena-text-to-image-snapshot-2026-10-07", "score": 1425, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-07", "sourceRow": "gpt-image-2.5-sunburst" }, { "benchmarkId": "arena-image-edit", "benchmarkSnapshotId": "arena-image-edit-snapshot-2026-10-06", "score": 1524, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-06", "sourceRow": "gpt-image-2.5-sunburst" } ], "officialSourceIds": [ "openai-gpt-image-2-5-sunburst-model", "openai-api-changelog" ] }, { "id": "gpt-image-2-5-flare", "providerId": "openai", "name": "GPT Image 2.5 Flare", "family": "GPT Image", "releaseDate": "2026-09-08", "status": "active", "access": "api", "apiModelId": "gpt-image-2.5-flare", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image" ], "taskFamilies": [ "image" ], "capabilities": [ "image-generation", "image-editing", "image-to-image" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "image" ], "strengths": [ "OpenAI's fastest model for high-quality everyday image generation, with six quality settings from low to max", "Token rates match GPT Image 2: $5 per 1M text input tokens, $8 per 1M image input tokens and $30 per 1M image output tokens, half that in batch" ], "cautions": [ "OpenAI publishes no per-image price, and its GPT Image 2 calculator does not estimate 2.5 token use, so cost per image has to be measured", "No independent image-quality benchmark is recorded here" ], "bestFor": [ "Fast, high-volume everyday image generation" ], "pricing": { "inputPerMillionUsd": 5, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": 1.25, "sourceId": "openai-gpt-image-2-5-flare-model" }, "prices": [ { "unit": "million-input-tokens", "usd": 5, "tier": "standard", "sourceId": "openai-gpt-image-2-5-flare-model" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "openai-gpt-image-2-5-flare-model" }, { "unit": "million-input-tokens", "usd": 1.25, "tier": "cached", "sourceId": "openai-gpt-image-2-5-flare-model" }, { "unit": "million-input-tokens", "usd": 8, "tier": "standard-image-input", "sourceId": "openai-gpt-image-2-5-flare-model" }, { "unit": "million-output-tokens", "usd": 30, "tier": "standard-image-output", "sourceId": "openai-gpt-image-2-5-flare-model" }, { "unit": "million-output-tokens", "usd": 15, "tier": "batch-image-output", "sourceId": "openai-gpt-image-2-5-flare-model" } ], "benchmarkResults": [ { "benchmarkId": "arena-text-to-image", "benchmarkSnapshotId": "arena-text-to-image-snapshot-2026-10-07", "score": 1398, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-07", "sourceRow": "gpt-image-2.5-flare" }, { "benchmarkId": "arena-image-edit", "benchmarkSnapshotId": "arena-image-edit-snapshot-2026-10-06", "score": 1481, "sourceId": "arena-leaderboard-dataset-2026-10-10", "measuredAt": "2026-10-06", "sourceRow": "gpt-image-2.5-flare" } ], "officialSourceIds": [ "openai-gpt-image-2-5-flare-model", "openai-api-changelog" ] }, { "id": "grok-imagine-video-1-5-lite", "providerId": "xai", "name": "Grok Imagine Video 1.5 Lite", "family": "Grok Imagine", "releaseDate": null, "status": "active", "access": "api", "apiModelId": "grok-imagine-video-1.5-lite", "contextWindowTokens": null, "maxOutputTokens": null, "modalities": [ "text", "image", "video" ], "taskFamilies": [ "video" ], "capabilities": [ "text-to-video", "image-to-video" ], "inputModalities": [ "text", "image" ], "outputModalities": [ "video" ], "strengths": [ "xAI's lowest-cost video model, generating from text or an image at $0.020 per output second", "Supports the batch API" ], "cautions": [ "xAI has not published a release date; trackers report early October 2026", "No independent video-quality benchmark is recorded here" ], "bestFor": [ "Cheap drafts and simple clips before committing to a higher-end video model" ], "pricing": { "inputPerMillionUsd": null, "outputPerMillionUsd": null, "cachedInputPerMillionUsd": null, "sourceId": "xai-grok-imagine-video-1-5-lite-model" }, "prices": [ { "unit": "million-input-tokens", "usd": null, "tier": "standard", "sourceId": "xai-grok-imagine-video-1-5-lite-model" }, { "unit": "million-output-tokens", "usd": null, "tier": "standard", "sourceId": "xai-grok-imagine-video-1-5-lite-model" }, { "unit": "million-input-tokens", "usd": null, "tier": "cached", "sourceId": "xai-grok-imagine-video-1-5-lite-model" }, { "unit": "video-second", "usd": 0.02, "tier": "standard", "sourceId": "xai-grok-imagine-video-1-5-lite-model" } ], "benchmarkResults": [], "officialSourceIds": [ "xai-grok-imagine-video-1-5-lite-model" ] } ], "sources": [ { "id": "livebench-2026-06-25-averaging-19e766a5", "title": "LiveBench category-weighted averaging implementation", "publisher": "LiveBench", "url": "https://github.com/LiveBench/livebench.github.io/blob/19e766a5de4de07d672ed5bf9f0a69ceed1d39bf/src/Table/Averaging.js", "sourceType": "independent", "verifiedAt": "2026-08-07", "datasetVersion": "LiveBench-2026-06-25@19e766a5de4de07d672ed5bf9f0a69ceed1d39bf", "license": "Apache-2.0", "topics": [ "benchmark" ], "reuseNote": "Pinned implementation averages task scores within each of the seven categories, then averages the category scores; the paired categories_2026_06_25.json file defines category membership." }, { "id": "livebench-2026-06-25-results-19e766a5", "title": "LiveBench 2026-06-25 task results", "publisher": "LiveBench", "url": "https://github.com/LiveBench/livebench.github.io/blob/19e766a5de4de07d672ed5bf9f0a69ceed1d39bf/public/table_2026_06_25.csv", "sourceType": "independent", "verifiedAt": "2026-08-07", "datasetVersion": "LiveBench-2026-06-25@19e766a5de4de07d672ed5bf9f0a69ceed1d39bf", "license": "Apache-2.0", "topics": [ "benchmark" ], "reuseNote": "Snapshot and measurement date: 2026-06-25. Overall scores are imported only for unambiguous model mappings and are recomputed with the pinned category-weighted averaging implementation." }, { "id": "openai-gpt-5-4-model", "title": "GPT-5.4 model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-5.4", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official model documentation verifies the dated snapshot, token limits, modalities, and exact standard token prices represented here." }, { "id": "openai-gpt-5-5-release", "title": "Introducing GPT-5.5", "publisher": "OpenAI", "url": "https://openai.com/index/introducing-gpt-5-5/", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official release announcement verifies model identity and the 2026-04-23 release date." }, { "id": "openai-gpt-5-5-model", "title": "GPT-5.5 model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-5.5", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official model documentation verifies token limits, modalities, and exact standard token prices represented here." }, { "id": "openai-gpt-5-6-release", "title": "GPT-5.6: Frontier intelligence that scales with your ambition", "publisher": "OpenAI", "url": "https://openai.com/index/gpt-5-6/", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official release announcement verifies the Sol, Terra, and Luna identities and 2026-07-09 general-availability date." }, { "id": "openai-gpt-5-6-models", "title": "OpenAI GPT-5.6 model catalog", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official model catalog verifies the Sol, Terra, and Luna token limits, modalities, roles, and exact standard token prices represented here." }, { "id": "anthropic-claude-opus-4-8-release", "title": "Introducing Claude Opus 4.8", "publisher": "Anthropic", "url": "https://www.anthropic.com/news/claude-opus-4-8", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official release documentation verifies model identity, release date, the published context window, and standard input/output pricing." }, { "id": "anthropic-claude-sonnet-5-release", "title": "Introducing Claude Sonnet 5", "publisher": "Anthropic", "url": "https://www.anthropic.com/news/claude-sonnet-5", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official release documentation verifies model identity, release date, API availability, and introductory pricing through 2026-08-31." }, { "id": "anthropic-claude-fable-5-release", "title": "Claude Fable 5 and Claude Mythos 5", "publisher": "Anthropic", "url": "https://www.anthropic.com/news/claude-fable-5-mythos-5", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official release announcement verifies Claude Fable 5 identity, 2026-06-09 release date, API access, standard token pricing, and general availability after redeployment." }, { "id": "anthropic-models-overview-current", "title": "Claude models overview", "publisher": "Anthropic", "url": "https://platform.claude.com/docs/en/about-claude/models/overview", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official current-model documentation verifies Fable's text-and-image inputs, 1,000,000-token context window, 128,000-token output limit, and current API availability." }, { "id": "google-gemini-3-5-changelog", "title": "Gemini API release notes", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/changelog", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Official release notes verify the dated Gemini 3.5 Flash, Gemini 3.1 Flash Live, Gemini 3.1 Flash TTS, Gemini 3.1 Flash Lite Image, Gemini 3 Pro Image, and Gemini Omni Flash releases represented in this catalog." }, { "id": "google-gemini-3-5-model", "title": "Gemini 3.5 Flash model", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.5-flash", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Official model documentation verifies model code, modalities, and token limits; exact token prices remain unreported here." }, { "id": "google-gemini-3-6-model", "title": "Gemini 3.6 Flash model", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.6-flash", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Official model documentation verifies model code, 2026-07-21 update, modalities, and token limits." }, { "id": "google-gemini-3-6-latest", "title": "Using the latest Gemini models", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/latest-model", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity", "pricing" ], "reuseNote": "Official latest-model guidance verifies the exact standard input/output prices represented for Gemini 3.6 Flash." }, { "id": "google-gemini-3-5-flash-lite-model", "title": "Gemini 3.5 Flash-Lite model", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.5-flash-lite", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Official model documentation verifies the stable model code, multimodal inputs, token limits, capabilities, and July 2026 update." }, { "id": "google-gemini-3-5-flash-lite-pricing", "title": "Gemini 3.5 Flash-Lite pricing", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/pricing#gemini-3.5-flash-lite", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "pricing" ], "reuseNote": "Official pricing table verifies the standard paid-tier input, cached-input, and output rates represented here." }, { "id": "xai-grok-4-3-model", "title": "Grok 4.3 model", "publisher": "SpaceXAI", "url": "https://docs.x.ai/developers/models/grok-4.3", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official model documentation verifies Grok 4.3 modalities, context window, reasoning modes, aliases, and standard short-context prices." }, { "id": "xai-inference-models-api", "title": "xAI inference models API reference", "publisher": "SpaceXAI", "url": "https://docs.x.ai/developers/rest-api-reference/inference/models", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official API reference exposes Grok 4.3 modalities, aliases, and token pricing fields; its object creation timestamp is not treated as a public release date." }, { "id": "xai-grok-4-3-public-availability", "title": "Skills in web, iOS, and Android", "publisher": "SpaceXAI", "url": "https://x.ai/news/grok-skills", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official 2026-05-18 announcement confirms Grok 4.3 was live across grok.com, iOS, and Android by that date; it is not treated as the model's exact release date." }, { "id": "xai-release-notes-current", "title": "SpaceXAI API release notes", "publisher": "SpaceXAI", "url": "https://docs.x.ai/developers/release-notes", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official release notes verify Grok 4.20 and Grok 4.20 Multi-agent API availability on 2026-03-10." }, { "id": "xai-grok-4-20-system-card", "title": "Grok 4.20 System Card", "publisher": "SpaceXAI", "url": "https://data.x.ai/2026-04-07-grok-4-20-model-card.pdf", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": "2026-04-07", "license": null, "topics": [ "identity" ], "reuseNote": "Official system card verifies Grok 4.20 identity, single-agent and multi-agent deployment modes, supported input modalities, and intended uses." }, { "id": "xai-current-model-pricing", "title": "SpaceXAI model pricing", "publisher": "SpaceXAI", "url": "https://docs.x.ai/developers/pricing", "sourceType": "official", "verifiedAt": "2026-08-14", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official current pricing table verifies active Grok 4.6, Grok 4.5, and Grok 4.20 variants, their documented context windows, and standard and long-context token rates." }, { "id": "xai-may-15-retirement", "title": "Grok model retirement on May 15, 2026", "publisher": "SpaceXAI", "url": "https://docs.x.ai/developers/migration/may-15-retirement", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official lifecycle notice verifies Grok 3 retirement and redirect to Grok 4.3, and identifies Grok 4.3 as the recommended general replacement." }, { "id": "xai-grok-4-5-release", "title": "Introducing Grok 4.5", "publisher": "SpaceXAI", "url": "https://x.ai/news/grok-4-5", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official announcement verifies the model identity, release date, and intended coding, agentic, and knowledge-work scope." }, { "id": "xai-grok-4-5-model", "title": "Grok 4.5 model", "publisher": "SpaceXAI", "url": "https://docs.x.ai/developers/models/grok-4.5", "sourceType": "official", "verifiedAt": "2026-08-07", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official model documentation verifies modalities, context window, and standard short-context token prices; the model caution discloses higher long-context rates." }, { "id": "xai-grok-4-6-release", "title": "Introducing Grok 4.6", "publisher": "SpaceXAI", "url": "https://x.ai/news/grok-4-6", "sourceType": "official", "verifiedAt": "2026-08-14", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official announcement verifies the 2026-08-12 release date and positioning for long-running agents, coding, and knowledge work." }, { "id": "xai-grok-4-6-model", "title": "Grok 4.6 model", "publisher": "SpaceXAI", "url": "https://docs.x.ai/developers/models/grok-4.6", "sourceType": "official", "verifiedAt": "2026-08-14", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official model documentation verifies the exact model ID, text and image input, text output, 500,000-token context window, reasoning, function calling, structured outputs, and standard token prices." }, { "id": "livebench-2024-11-25-methodology", "title": "LiveBench datasheet and methodology", "publisher": "LiveBench", "url": "https://github.com/LiveBench/LiveBench/blob/00eae856aa1c1a9e9d058a65a9a94d85884034c4/docs/DATASHEET.md", "sourceType": "independent", "verifiedAt": "2026-08-06", "datasetVersion": "LiveBench-2024-11-25", "license": "Apache-2.0", "topics": [ "benchmark" ], "reuseNote": "Primary benchmark documentation confirms objective scoring, public distribution, and Apache 2.0 reuse terms." }, { "id": "livebench-2024-11-25-results-347e3d0b", "title": "LiveBench 2024-11-25 task results", "publisher": "LiveBench", "url": "https://github.com/LiveBench/livebench.github.io/blob/347e3d0b6a4916cd66f8bda31ce7ecdd7436ecc3/public/table_2024_11_25.csv", "sourceType": "independent", "verifiedAt": "2026-08-06", "datasetVersion": "LiveBench-2024-11-25@347e3d0b6a4916cd66f8bda31ce7ecdd7436ecc3", "license": "Apache-2.0", "topics": [ "benchmark" ], "reuseNote": "Snapshot and measurement date: 2024-11-25. Exact published snapshot; only direct Web of Lies V2 percentages with unambiguous model mappings are imported." }, { "id": "openai-gpt-4o-release", "title": "Hello GPT-4o", "publisher": "OpenAI", "url": "https://openai.com/index/hello-gpt-4o/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "openai-gpt-4-1-release", "title": "Introducing GPT-4.1 in the API", "publisher": "OpenAI", "url": "https://openai.com/index/gpt-4-1/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "openai-pricing", "title": "OpenAI API pricing", "publisher": "OpenAI", "url": "https://openai.com/api/pricing/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "pricing" ], "reuseNote": "Official pricing page used only for the exact token prices represented in model records." }, { "id": "anthropic-claude-3-release", "title": "Introducing the next generation of Claude", "publisher": "Anthropic", "url": "https://www.anthropic.com/news/claude-3-family", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "anthropic-claude-3-7-release", "title": "Claude 3.7 Sonnet and Claude Code", "publisher": "Anthropic", "url": "https://www.anthropic.com/news/claude-3-7-sonnet", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "anthropic-pricing", "title": "Claude pricing", "publisher": "Anthropic", "url": "https://docs.anthropic.com/en/docs/about-claude/pricing", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "pricing" ], "reuseNote": "Official pricing documentation used only for the exact token prices represented in model records." }, { "id": "google-gemini-1-5-release", "title": "Our next-generation model: Gemini 1.5", "publisher": "Google", "url": "https://blog.google/technology/ai/google-gemini-next-generation-model-february-2024/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "google-gemini-1-5-flash-release", "title": "Gemini 1.5 Flash", "publisher": "Google", "url": "https://developers.googleblog.com/en/gemini-15-flash-updates-google-ai-studio-gemini-api/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "google-gemini-2-release", "title": "Introducing Gemini 2.0", "publisher": "Google", "url": "https://blog.google/technology/google-deepmind/google-gemini-ai-update-december-2024/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "google-gemini-2-5-release", "title": "Gemini 2.5: Our most intelligent AI model", "publisher": "Google", "url": "https://blog.google/technology/google-deepmind/gemini-model-thinking-updates-march-2025/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "google-pricing", "title": "Gemini Developer API pricing", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/pricing", "sourceType": "official", "verifiedAt": "2026-10-11", "datasetVersion": null, "license": null, "topics": [ "pricing" ], "reuseNote": "Official Gemini Developer API pricing table used only for exact published token, image, audio-minute, and video-second prices represented in model records." }, { "id": "meta-llama-3-release", "title": "Introducing Meta Llama 3", "publisher": "Meta", "url": "https://ai.meta.com/blog/meta-llama-3/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "meta-llama-3-1-release", "title": "Introducing Llama 3.1", "publisher": "Meta", "url": "https://ai.meta.com/blog/meta-llama-3-1/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "meta-llama-3-2-release", "title": "Llama 3.2", "publisher": "Meta", "url": "https://ai.meta.com/blog/llama-3-2-connect-2024-vision-edge-mobile-devices/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "xai-grok-1-5-release", "title": "Announcing Grok-1.5", "publisher": "xAI", "url": "https://x.ai/news/grok-1.5", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "xai-grok-1-5v-release", "title": "Grok-1.5 Vision Preview", "publisher": "xAI", "url": "https://x.ai/news/grok-1.5v", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "xai-grok-2-release", "title": "Grok-2 Beta Release", "publisher": "xAI", "url": "https://x.ai/news/grok-2", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "xai-grok-3-release", "title": "Grok 3 Beta", "publisher": "xAI", "url": "https://x.ai/news/grok-3", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "xai-models-pricing", "title": "xAI models and pricing", "publisher": "xAI", "url": "https://docs.x.ai/docs/models", "sourceType": "official", "verifiedAt": "2026-10-11", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official model catalog used to verify documented model availability and pricing fields, not historical release dates." }, { "id": "deepseek-v2-release", "title": "DeepSeek-V2", "publisher": "DeepSeek", "url": "https://github.com/deepseek-ai/DeepSeek-V2", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "deepseek-coder-v2-release", "title": "DeepSeek-Coder-V2", "publisher": "DeepSeek", "url": "https://github.com/deepseek-ai/DeepSeek-Coder-V2", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "deepseek-v3-release", "title": "DeepSeek-V3", "publisher": "DeepSeek", "url": "https://github.com/deepseek-ai/DeepSeek-V3", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "deepseek-r1-release", "title": "DeepSeek-R1", "publisher": "DeepSeek", "url": "https://github.com/deepseek-ai/DeepSeek-R1", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "deepseek-pricing", "title": "DeepSeek API pricing", "publisher": "DeepSeek", "url": "https://api-docs.deepseek.com/quick_start/pricing", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "pricing" ], "reuseNote": "Official pricing documentation consulted for price provenance; unavailable exact values remain unreported." }, { "id": "mistral-codestral-release", "title": "Codestral", "publisher": "Mistral AI", "url": "https://mistral.ai/news/codestral/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "mistral-large-2-release", "title": "Large Enough", "publisher": "Mistral AI", "url": "https://mistral.ai/news/mistral-large-2407/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "mistral-pixtral-release", "title": "Announcing Pixtral 12B", "publisher": "Mistral AI", "url": "https://mistral.ai/news/pixtral-12b/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "mistral-ministral-release", "title": "Un Ministral, des Ministraux", "publisher": "Mistral AI", "url": "https://mistral.ai/news/ministraux/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "mistral-pricing", "title": "Mistral AI pricing", "publisher": "Mistral AI", "url": "https://mistral.ai/products/la-plateforme#pricing", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "pricing" ], "reuseNote": "Official pricing page consulted for price provenance; unavailable exact values remain unreported." }, { "id": "qwen2-release", "title": "Hello Qwen2", "publisher": "Qwen Team", "url": "https://qwenlm.github.io/blog/qwen2/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "qwen2-5-release", "title": "Qwen2.5: A Party of Foundation Models", "publisher": "Qwen Team", "url": "https://qwenlm.github.io/blog/qwen2.5/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "qwen2-5-vl-release", "title": "Qwen2.5-VL", "publisher": "Qwen Team", "url": "https://qwenlm.github.io/blog/qwen2.5-vl/", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "kimi-vl-release", "title": "Kimi-VL", "publisher": "Moonshot AI", "url": "https://github.com/MoonshotAI/Kimi-VL", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "kimi-dev-release", "title": "Kimi-Dev", "publisher": "Moonshot AI", "url": "https://github.com/MoonshotAI/Kimi-Dev", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "kimi-k2-release", "title": "Kimi K2", "publisher": "Moonshot AI", "url": "https://github.com/MoonshotAI/Kimi-K2", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify model identity and release details." }, { "id": "openai-gpt-4o-model", "title": "GPT-4o model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-4o", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official provider documentation used to verify exact model specifications." }, { "id": "openai-gpt-4-1-model", "title": "GPT-4.1 model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-4.1", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official provider documentation used to verify exact model specifications." }, { "id": "openai-gpt-4-1-mini-model", "title": "GPT-4.1 mini model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-4.1-mini", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official provider documentation used to verify exact model specifications." }, { "id": "openai-gpt-4-1-nano-model", "title": "GPT-4.1 nano model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-4.1-nano", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official provider documentation used to verify exact model specifications." }, { "id": "anthropic-claude-3-haiku-release", "title": "Claude 3 Haiku", "publisher": "Anthropic", "url": "https://www.anthropic.com/news/claude-3-haiku", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify exact model specifications." }, { "id": "anthropic-models-overview", "title": "Claude models overview", "publisher": "Anthropic", "url": "https://docs.anthropic.com/en/docs/about-claude/models/overview", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify exact model specifications." }, { "id": "google-models-overview", "title": "Gemini API models", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify exact model specifications." }, { "id": "google-gemini-2-5-pro-model", "title": "Gemini 2.5 Pro", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-2.5-pro", "sourceType": "official", "verifiedAt": "2026-08-06", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official provider documentation used to verify exact model specifications." }, { "id": "specialist-openai-gpt-image-2", "title": "GPT Image 2 model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-image-2", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-openai-realtime-1-5", "title": "GPT-Realtime-1.5 model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-realtime-1.5", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-openai-audio-1-5", "title": "gpt-audio-1.5 model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-audio-1.5", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-openai-transcribe", "title": "GPT-4o Transcribe model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-4o-transcribe", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-openai-tts-1", "title": "TTS-1 model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/tts-1", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-google-gemini-3-1-flash-image-pricing", "title": "Gemini 3.1 Flash Image pricing", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/pricing#gemini-3.1-flash-image-nano-banana-2", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party pricing table used to verify Gemini 3.1 Flash Image standard token and resolution-specific image prices." }, { "id": "specialist-google-native-image-capabilities", "title": "Gemini native image generation capabilities", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/image-generation", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "First-party native image guide used to verify Gemini 3.1 Flash Image generation, editing, image input, and mixed text/image output capabilities." }, { "id": "specialist-google-native-media", "title": "Gemini Developer API pricing and native media model identifiers", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/pricing#veo-3.1", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-xai-imagine-image", "title": "grok-imagine-image model", "publisher": "SpaceXAI", "url": "https://docs.x.ai/developers/models/grok-imagine-image", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-xai-imagine-video", "title": "Video generation", "publisher": "SpaceXAI", "url": "https://docs.x.ai/developers/model-capabilities/video/generation", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-xai-imagine-pricing", "title": "Imagine API pricing", "publisher": "SpaceXAI", "url": "https://docs.x.ai/developers/pricing#imagine-api", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-runway-models", "title": "Available AI Models", "publisher": "Runway", "url": "https://docs.dev.runwayml.com/guides/models/", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-elevenlabs-models", "title": "ElevenLabs models", "publisher": "ElevenLabs", "url": "https://elevenlabs.io/docs/overview/models", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-elevenlabs-scribe-v2-release", "title": "Introducing Scribe v2", "publisher": "ElevenLabs", "url": "https://elevenlabs.io/blog/introducing-scribe-v2", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-elevenlabs-scribe-realtime-release", "title": "Introducing Scribe v2 Realtime", "publisher": "ElevenLabs", "url": "https://elevenlabs.io/blog/introducing-scribe-v2-realtime", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-stability-pricing", "title": "Stability AI Developer Platform pricing", "publisher": "Stability AI", "url": "https://platform.stability.ai/pricing", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-bfl-models", "title": "FLUX image generation endpoints", "publisher": "Black Forest Labs", "url": "https://docs.bfl.ai/quick_start/generating_images", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "specialist-bfl-pricing", "title": "FLUX API pricing", "publisher": "Black Forest Labs", "url": "https://docs.bfl.ai/quick_start/pricing", "sourceType": "official", "verifiedAt": "2026-08-10", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "anthropic-pricing-2026-09", "title": "Claude API pricing", "publisher": "Anthropic", "url": "https://platform.claude.com/docs/en/about-claude/pricing", "sourceType": "official", "verifiedAt": "2026-10-11", "datasetVersion": null, "license": null, "topics": [ "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "anthropic-models-overview-2026-09", "title": "Claude models overview", "publisher": "Anthropic", "url": "https://platform.claude.com/docs/en/models/overview", "sourceType": "official", "verifiedAt": "2026-10-11", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "anthropic-glasswing-2026-09", "title": "Claude Mythos limited availability", "publisher": "Anthropic", "url": "https://www.anthropic.com/glasswing", "sourceType": "official", "verifiedAt": "2026-09-09", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "openai-gpt-6-astra-model", "title": "GPT-6 Astra model reference", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-6-astra", "sourceType": "official", "verifiedAt": "2026-09-09", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "openai-api-pricing-2026-09", "title": "OpenAI API pricing", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/pricing", "sourceType": "official", "verifiedAt": "2026-10-11", "datasetVersion": null, "license": null, "topics": [ "pricing" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "google-gemini-38-flash-2026-09", "title": "Gemini 3.8 Flash model card", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash", "sourceType": "official", "verifiedAt": "2026-09-09", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "meta-model-api-models-2026-09", "title": "Meta Model API model list", "publisher": "Meta", "url": "https://dev.meta.ai/docs/models", "sourceType": "official", "verifiedAt": "2026-09-09", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "kimi-platform-pricing-2026-09", "title": "Kimi platform model and pricing documentation", "publisher": "Moonshot AI", "url": "https://platform.kimi.ai/docs/pricing/chat", "sourceType": "official", "verifiedAt": "2026-09-09", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "alibaba-model-studio-models-2026-09", "title": "Alibaba Cloud Model Studio model list", "publisher": "Alibaba Cloud", "url": "https://www.alibabacloud.com/help/en/model-studio/models", "sourceType": "official", "verifiedAt": "2026-09-09", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "First-party provider documentation used only for the stated model identity, capabilities, lifecycle, access, and pricing." }, { "id": "openrouter-models-api-2026-09", "title": "OpenRouter models API catalogue", "publisher": "OpenRouter", "url": "https://openrouter.ai/api/v1/models", "sourceType": "independent", "verifiedAt": "2026-09-09", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Third-party API catalogue used only to cross-check published token prices, context limits, and listing dates against first-party documentation." }, { "id": "google-gemini-3-7-flash-model", "title": "Gemini 3.7 Flash model", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.7-flash", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Official model documentation verifies the stable Gemini 3.7 Flash code, text/image/video/audio inputs, text output, token limits, structured outputs, thinking, and URL context." }, { "id": "google-gemini-3-7-flash-release", "title": "Introducing Gemini 3.7 Flash", "publisher": "Google", "url": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Official Google announcement verifies the 2026-08-13 release, coding and agent positioning, and introductory token prices through 2026-12-31 before the documented 2027 rate change." }, { "id": "google-gemini-3-1-pro-preview-model", "title": "Gemini 3.1 Pro Preview", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Model card verifies the Gemini 3.1 Pro Preview endpoint, multimodal inputs, text output, token limits, and documented capabilities." }, { "id": "google-gemini-3-flash-preview-model", "title": "Gemini 3 Flash Preview", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3-flash-preview", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Model card verifies the Gemini 3 Flash Preview endpoint, multimodal inputs, text output, token limits, and documented capabilities." }, { "id": "google-gemini-3-1-flash-lite-model", "title": "Gemini 3.1 Flash-Lite", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.1-flash-lite", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Model card verifies the stable Gemini 3.1 Flash-Lite endpoint, multimodal inputs, text output, token limits, and documented capabilities." }, { "id": "google-gemini-3-5-live-translate-model", "title": "Gemini 3.5 Live Translate", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.5-live-translate-preview", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Model card verifies the Gemini 3.5 Live Translate preview endpoint, audio input, translated-audio and transcript output, and token limits." }, { "id": "google-live-api-guide", "title": "Gemini Live API", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/live-api", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Live API guide verifies realtime audio model operation and the preview endpoint lifecycle applies to the live-model records." }, { "id": "google-gemini-3-1-flash-live-model", "title": "Gemini 3.1 Flash Live Preview", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.1-flash-live-preview", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Model card verifies the Gemini 3.1 Flash Live preview endpoint, supported realtime multimodal inputs, text and audio output, token limits, and capabilities." }, { "id": "google-gemini-3-1-flash-tts-model", "title": "Gemini 3.1 Flash TTS Preview", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.1-flash-tts-preview", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Model card verifies the Gemini 3.1 Flash TTS preview endpoint, text input, audio output, token limits, and audio generation capability." }, { "id": "google-tts-guide", "title": "Text-to-speech generation", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/speech-generation", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Text-to-Speech guide verifies controllable speech generation and streaming support for the Gemini 3.1 Flash TTS preview endpoint." }, { "id": "google-gemini-3-1-flash-lite-image-model", "title": "Gemini 3.1 Flash Lite Image", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.1-flash-lite-image", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Model card verifies the stable Gemini 3.1 Flash Lite Image endpoint, text and image inputs, text and image output, token limits, and image generation and editing capabilities." }, { "id": "google-gemini-3-pro-image-model", "title": "Gemini 3 Pro Image", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3-pro-image", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Model card verifies the stable Gemini 3 Pro Image endpoint, text and image inputs and outputs, token limits, and image generation capability." }, { "id": "google-gemini-omni-flash-model", "title": "Gemini Omni Flash", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-omni-flash", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Model card verifies the Gemini Omni Flash preview endpoint, text, image, and video inputs, video output, context window, and 3-to-10-second output range." }, { "id": "google-omni-flash-guide", "title": "Generate and edit videos with Gemini Omni Flash", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/omni", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Gemini Omni Flash guide verifies preview lifecycle, conversational video editing, native multimodal context, and generated video with audio." }, { "id": "google-gemini-deprecations", "title": "Gemini deprecations", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/deprecations", "sourceType": "official", "verifiedAt": "2026-08-18", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "identity" ], "reuseNote": "Lifecycle table verifies published stable or preview status and announced shutdown dates only for the Gemini endpoints that reference it." }, { "id": "epoch-benchmark-data-2026-10-12", "title": "Epoch AI Benchmarking Hub data (benchmark_data.zip)", "publisher": "Epoch AI", "url": "https://epoch.ai/benchmarks/use-this-data", "sourceType": "independent", "verifiedAt": "2026-10-12", "datasetVersion": "benchmark_data.zip retrieved 2026-10-12, sha256 prefix 2fb16ca26aea846c", "license": "CC-BY-4.0", "topics": [ "benchmark" ], "reuseNote": "Results Epoch AI ran itself (the GPQA Diamond, FrontierMath Tiers 1-3 v2 and SimpleQA Verified files), not the external files it collects from other publishers. A model maps only when its Epoch row is unambiguous: exact name, or a dated version pinned by the record release date. Where Epoch lists several reasoning-effort settings, the best published configuration is used and its row is recorded in sourceRow. Scores are converted from 0-1 to 0-100 and rounded to one decimal. Epoch evaluated rows between 2023 and 2026-10-09; every result is dated to this capture. Credit: Epoch AI, Capabilities and benchmarking, epoch.ai, CC BY 4.0." }, { "id": "epoch-gpqa-diamond", "title": "GPQA Diamond benchmark (Epoch AI)", "publisher": "Epoch AI", "url": "https://epoch.ai/benchmarks/gpqa-diamond", "sourceType": "independent", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "benchmark" ], "reuseNote": "198 four-option graduate-level questions in biology, physics and chemistry, written by domain experts; random guessing scores 25%. Epoch runs the evaluation itself." }, { "id": "epoch-frontiermath", "title": "FrontierMath benchmark (Epoch AI)", "publisher": "Epoch AI", "url": "https://epoch.ai/benchmarks/frontiermath", "sourceType": "independent", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "benchmark" ], "reuseNote": "Original research-level mathematics problems written and vetted by expert mathematicians. Tiers 1-3 v2 (released 2026-06-12) is the 295-problem base set after a revision that corrected errors in 42% of problems. Scores come from Epoch-run private sets. OpenAI funded the benchmark and has access to part of it." }, { "id": "epoch-simpleqa-verified", "title": "SimpleQA Verified benchmark (Epoch AI)", "publisher": "Epoch AI", "url": "https://epoch.ai/benchmarks/simpleqa-verified", "sourceType": "independent", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "benchmark" ], "reuseNote": "A 1,000-question factoid benchmark from Google, building on SimpleQA from OpenAI. Epoch runs it itself and reports the proportion answered correctly, not the F1-based score Google reports." }, { "id": "deepseek-v4-1-flash-release", "title": "DeepSeek-V4.1-Flash release", "publisher": "DeepSeek", "url": "https://api-docs.deepseek.com/news/news260910", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official release note, 2026-09-10: 552B-parameter MoE with native visual understanding; 8B active parameters for input and 16B for output." }, { "id": "deepseek-v4-1-flash-weights", "title": "DeepSeek-V4.1-Flash model weights", "publisher": "DeepSeek", "url": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official open-weights repository, created 2026-09-10, MIT licence, image-text-to-text." }, { "id": "mistral-large-4-release", "title": "Introducing Mistral Large 4", "publisher": "Mistral AI", "url": "https://mistral.ai/news/mistral-large-4/", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Public preview announced 2026-10-06 at $1.36 / $4.18 per 1M input / output tokens; weights promised by the end of October." }, { "id": "mistral-large-4-model", "title": "Mistral Large 4 model card", "publisher": "Mistral AI", "url": "https://docs.mistral.ai/models/mistral-large-4-0", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Model id mistral-large-4-0, public preview, 1M context, 1.05T total and 52B active parameters, 1.6B vision encoder." }, { "id": "qwen-3-8-omni-flash-model", "title": "qwen3.8-omni-flash model information", "publisher": "Alibaba Cloud", "url": "https://www.alibabacloud.com/help/en/model-studio/qwen3-8-omni-flash", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Text, image, audio and video input; text output; 1M-token context; 131,072 max output tokens." }, { "id": "qwen-3-8-omni-flash-blog", "title": "Qwen3.8-Omni-Flash: Omni Senses. Agentic Delivery.", "publisher": "Alibaba Cloud", "url": "https://www.alibabacloud.com/blog/qwen3-8-omni-flash-omni-senses--agentic-delivery-_603580", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Official announcement published 2026-09-20." }, { "id": "alibaba-model-studio-pricing", "title": "Alibaba Cloud Model Studio model pricing", "publisher": "Alibaba Cloud", "url": "https://www.alibabacloud.com/help/en/model-studio/model-pricing", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "pricing" ], "reuseNote": "International (Singapore) rates used." }, { "id": "google-nano-banana-2-1-model", "title": "Gemini Nano Banana 2.1", "publisher": "Google", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-nano-banana-2.1", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Stable model gemini-nano-banana-2.1: text, image, video and PDF input; image and text output; 1K, 2K and 4K output." }, { "id": "google-gemini-3-8-live-release", "title": "Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking", "publisher": "Google", "url": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-live-gemini-3-8-live-extended-thinking/", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Two speech-to-speech models for the Gemini API and Google AI Studio." }, { "id": "google-gemini-3-8-tts-release", "title": "Gemini 3.8 text-to-speech models", "publisher": "Google", "url": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-text-to-speech/", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Gemini 3.8 Flash TTS and Flash-Lite TTS announcement." }, { "id": "openai-gpt-live-1-model", "title": "GPT-Live 1 model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-live-1", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Full-duplex voice model; audio and text in and out; $0.05 per voice-session minute billed per second; reasoning and tools delegated to a separately billed backend model." }, { "id": "openai-api-changelog", "title": "OpenAI API changelog", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/changelog", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity" ], "reuseNote": "Release dates: gpt-image-2.5-sunburst and gpt-image-2.5-flare 2026-09-08; gpt-live-1 generally available 2026-09-10." }, { "id": "openai-gpt-image-2-5-sunburst-model", "title": "GPT Image 2.5 Sunburst model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-image-2.5-sunburst", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Most capable image generation and editing model; token rates match GPT Image 2." }, { "id": "openai-gpt-image-2-5-flare-model", "title": "GPT Image 2.5 Flare model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-image-2.5-flare", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Fastest model for everyday image generation; token rates match GPT Image 2." }, { "id": "xai-grok-imagine-video-1-5-lite-model", "title": "grok-imagine-video-1.5-lite", "publisher": "xAI", "url": "https://docs.x.ai/developers/models/grok-imagine-video-1.5-lite", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "Text and image to video at $0.020 per output second; batch API supported." }, { "id": "openai-gpt-5-6-sol-model", "title": "GPT-5.6 Sol model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-5.6-sol", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "1.05M context, 128K max output, knowledge cutoff 2026-02-16, six reasoning-effort levels." }, { "id": "openai-gpt-5-6-terra-model", "title": "GPT-5.6 Terra model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-5.6-terra", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "1.05M context, 128K max output, knowledge cutoff 2026-02-16, six reasoning-effort levels." }, { "id": "openai-gpt-5-6-luna-model", "title": "GPT-5.6 Luna model", "publisher": "OpenAI", "url": "https://developers.openai.com/api/docs/models/gpt-5.6-luna", "sourceType": "official", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": null, "topics": [ "identity", "pricing" ], "reuseNote": "1.05M context, 128K max output, knowledge cutoff 2026-02-16, six reasoning-effort levels." }, { "id": "arena-leaderboard-dataset-2026-10-10", "title": "Arena leaderboard dataset (lmarena-ai/leaderboard-dataset)", "publisher": "Arena", "url": "https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/tree/e16f9023298691b4596eca61b4e31161dff31200", "sourceType": "independent", "verifiedAt": "2026-10-12", "datasetVersion": "lmarena-ai/leaderboard-dataset@e16f9023298691b4596eca61b4e31161dff31200", "license": "CC-BY-4.0", "topics": [ "benchmark" ], "reuseNote": "Ratings from blind pairwise human votes, published by Arena (formerly LMArena) in its public leaderboard dataset and pinned here to one commit. The overall category of the latest split is used; for text, the style-controlled leaderboard, which is the default view on arena.ai. A model maps only when its Arena row names the same model; where Arena lists several quality or resolution settings of one model, the best rated row is used and recorded in sourceRow. Ratings are rounded to whole points. Credit: Arena, lmarena-ai/leaderboard-dataset on Hugging Face, CC BY 4.0." }, { "id": "arena-text-to-image", "title": "Arena Text-to-Image leaderboard", "publisher": "Arena", "url": "https://arena.ai/leaderboard/image/text-to-image", "sourceType": "independent", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "benchmark" ], "reuseNote": "People compare two anonymous images generated from the same prompt and vote for the better one. Arena fits a Bradley-Terry model to the votes and reports ratings on an Elo-like scale." }, { "id": "arena-image-edit", "title": "Arena Image Edit leaderboard", "publisher": "Arena", "url": "https://arena.ai/leaderboard/image/image-edit", "sourceType": "independent", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "benchmark" ], "reuseNote": "People compare two anonymous edits of the same image from the same instruction and vote for the better one. Arena fits a Bradley-Terry model to the votes and reports ratings on an Elo-like scale." }, { "id": "arena-text-to-video", "title": "Arena Text-to-Video leaderboard", "publisher": "Arena", "url": "https://arena.ai/leaderboard/video/text-to-video", "sourceType": "independent", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "benchmark" ], "reuseNote": "People compare two anonymous videos generated from the same prompt and vote for the better one. Arena fits a Bradley-Terry model to the votes and reports ratings on an Elo-like scale." }, { "id": "arena-image-to-video", "title": "Arena Image-to-Video leaderboard", "publisher": "Arena", "url": "https://arena.ai/leaderboard/video/image-to-video", "sourceType": "independent", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "benchmark" ], "reuseNote": "People compare two anonymous videos animated from the same image and prompt and vote for the better one. Arena fits a Bradley-Terry model to the votes and reports ratings on an Elo-like scale." }, { "id": "arena-text", "title": "Arena Text leaderboard", "publisher": "Arena", "url": "https://arena.ai/leaderboard/chat/text", "sourceType": "independent", "verifiedAt": "2026-10-12", "datasetVersion": null, "license": "CC-BY-4.0", "topics": [ "benchmark" ], "reuseNote": "People put the same prompt to two anonymous models and vote for the better answer. Arena fits a Bradley-Terry model to the votes; the rating used here is the style-controlled one, which is the default view on arena.ai." } ], "recommendations": [ { "useCase": "agents", "modelIds": [ "gpt-56-sol", "claude-sonnet-5", "gemini-36-flash" ], "summary": "Start a current agent-workflow evaluation with GPT-5.6 Sol, Claude Sonnet 5, and Gemini 3.6 Flash.", "rationale": [ "All three active API records are positioned by their providers for agentic or tool-using workflows.", "GPT-5.6 Sol and Gemini 3.6 Flash publish context windows above one million tokens, while Claude Sonnet 5 launches with a lower introductory input price than GPT-5.6 Sol." ], "conditions": [ "Test the models with the actual tools, permissions, retry rules, and human approval steps used in production." ], "limitations": [ "The pinned LiveBench snapshot predates GPT-5.6 Sol and Gemini 3.6 Flash and does not measure tool-use success, latency, uptime, or end-to-end agent reliability; Claude Sonnet 5 pricing changes after 2026-08-31." ], "sourceIds": [ "openai-gpt-5-6-release", "openai-gpt-5-6-models", "anthropic-claude-sonnet-5-release", "google-gemini-3-6-model", "google-gemini-3-6-latest" ] }, { "useCase": "coding", "modelIds": [ "gpt-56-sol", "claude-fable-5" ], "summary": "Test GPT-5.6 Sol and Claude Fable 5 as current coding-focused API options before choosing a model for repository work.", "rationale": [ "Both providers position these active records for complex coding or software-engineering workflows.", "Their published million-token-class context windows support evaluation on large repositories and long-running work." ], "conditions": [ "Evaluate repository navigation, code edits, test execution, and review quality on your own languages and tooling." ], "limitations": [ "The catalog does not include a compatible repository-level coding benchmark for either model, and the pinned LiveBench snapshot predates both models." ], "sourceIds": [ "openai-gpt-5-6-release", "openai-gpt-5-6-models", "anthropic-claude-fable-5-release", "anthropic-models-overview-current" ] }, { "useCase": "analysis", "modelIds": [ "gpt-55", "claude-fable-5" ], "summary": "Compare GPT-5.5 and Claude Fable 5 for current high-capability analysis, then test both against your review standard.", "rationale": [ "Both providers position these active API records for demanding reasoning or professional knowledge work.", "Both records publish context windows of at least 1,000,000 tokens." ], "conditions": [ "Use representative documents, expected citations, and a defined human review rubric for the production trial." ], "limitations": [ "The pinned LiveBench snapshot does not provide a compatible Claude Fable 5 result and does not establish factuality, domain expertise, citation quality, latency, or production reliability." ], "sourceIds": [ "openai-gpt-5-5-model", "anthropic-claude-fable-5-release", "anthropic-models-overview-current" ] }, { "useCase": "multimodal", "modelIds": [ "gpt-56-terra", "gemini-36-flash", "grok-45" ], "summary": "Compare current API options from OpenAI, Google, and SpaceXAI when the workload combines text with visual inputs.", "rationale": [ "Gemini 3.6 Flash records text, image, audio, and video inputs; GPT-5.6 Terra and Grok 4.5 record text and image inputs.", "The three current records span published context windows from 500,000 to 1,050,000 tokens and different input-price points." ], "conditions": [ "Build an evaluation set from the image, audio, or video formats and document layouts the system will receive." ], "limitations": [ "The dataset has no compatible multimodal quality benchmark for this shortlist and does not record modality-specific pricing." ], "sourceIds": [ "openai-gpt-5-6-models", "google-gemini-3-6-model", "google-gemini-3-6-latest", "xai-grok-4-5-model" ] }, { "useCase": "budget", "modelIds": [ "gemini-35-flash-lite", "gpt-41-nano" ], "summary": "Compare current Gemini 3.5 Flash-Lite with GPT-4.1 nano for cost-sensitive API trials using published list prices.", "rationale": [ "Gemini 3.5 Flash-Lite records $0.30 input and $2.50 output per million tokens, while GPT-4.1 nano records $0.10 input and $0.40 output.", "Both active records accept multimodal input and publish context windows above one million tokens." ], "conditions": [ "Estimate full task cost with representative input size, output size, retries, caching, review, and supporting infrastructure." ], "limitations": [ "List price does not include every production cost, and this dataset has no compatible quality score for either record." ], "sourceIds": [ "google-gemini-3-5-flash-lite-model", "google-gemini-3-5-flash-lite-pricing", "openai-gpt-4-1-nano-model", "openai-pricing" ] } ] }