{"source":"AIMap","schema":"benchmark/v1","updated":"2026-10-22","benchmark":{"id":"math-level-5","name":"MATH Level 5","category":"math","unit":"%","higherIsBetter":true,"status":"saturated","sources":["epoch"],"primaryVariant":"epoch","url":"https://github.com/hendrycks/math","year":2021,"desc":{"en":"The hardest (level 5) problems of the MATH competition dataset (Hendrycks et al., 2021).","zh-hans":"MATH 竞赛数据集（Hendrycks 等，2021）中最难的第 5 级题目。"},"measures":{"en":"Accuracy, Epoch AI's own runs.","zh-hans":"准确率，Epoch AI 统一评测。"},"limits":{"en":"Near saturation for reasoning models.","zh-hans":"推理模型已接近满分。"},"contamination":"high","contaminationNote":{"en":"Public since 2021.","zh-hans":"2021 年起公开。"},"stats":{"results":104,"models":94,"latest":"2025-10-30","bySourceType":{"independent":104}}},"sources":[{"id":"epoch","name":"Epoch AI Benchmarking Hub","license":"CC BY 4.0","policy":"open","attribution":"Epoch AI, 'Capabilities & benchmarking', https://epoch.ai/benchmarks (CC BY 4.0)","site":"https://epoch.ai/benchmarks"}],"results":[{"model":"gpt-5","modelName":"GPT-5","provider":"OpenAI","rawName":"GPT-5","variant":"epoch","setting":"Epoch AI run, accuracy; effort: high","score":98.13,"unit":"%","rank":1,"rankComputedBySite":true,"date":"2025-10-29","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-5","modelName":"GPT-5","provider":"OpenAI","rawName":"GPT-5","variant":"epoch","setting":"Epoch AI run, accuracy; effort: medium","score":97.92,"unit":"%","rank":2,"rankComputedBySite":true,"date":"2025-08-20","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-5-mini","modelName":"GPT-5 mini","provider":"OpenAI","rawName":"GPT-5 mini","variant":"epoch","setting":"Epoch AI run, accuracy; effort: high","score":97.85,"unit":"%","rank":3,"rankComputedBySite":true,"date":"2025-10-30","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"o4-mini","modelName":"o4-mini","provider":"OpenAI","rawName":"o4-mini","variant":"epoch","setting":"Epoch AI run, accuracy; effort: high","score":97.83,"unit":"%","rank":4,"rankComputedBySite":true,"date":"2025-04-16","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"o3","modelName":"o3","provider":"OpenAI","rawName":"o3","variant":"epoch","setting":"Epoch AI run, accuracy; effort: high","score":97.77,"unit":"%","rank":5,"rankComputedBySite":true,"date":"2025-04-16","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-sonnet-4-5","modelName":"Claude Sonnet 4.5","provider":"Anthropic","rawName":"Claude Sonnet 4.5","variant":"epoch","setting":"Epoch AI run, accuracy; effort: 32k thinking","score":97.73,"unit":"%","rank":6,"rankComputedBySite":true,"date":"2025-10-21","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"qwen3-max","modelName":"Qwen3-Max","provider":"Alibaba (Qwen)","rawName":"Qwen3-Max","variant":"epoch","setting":"Epoch AI run, accuracy","score":97.13,"unit":"%","rank":7,"rankComputedBySite":true,"date":"2025-10-09","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-5-mini","modelName":"GPT-5 mini","provider":"OpenAI","rawName":"GPT-5 mini","variant":"epoch","setting":"Epoch AI run, accuracy; effort: medium","score":96.79,"unit":"%","rank":8,"rankComputedBySite":true,"date":"2025-08-20","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"deepseek-r1-may-2025","modelName":"DeepSeek-R1 (May 2025)","provider":"DeepSeek","rawName":"DeepSeek-R1 (May 2025)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: May 2025","score":96.64,"unit":"%","rank":9,"rankComputedBySite":true,"date":"2025-05-29","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"o3-mini","modelName":"o3-mini","provider":"OpenAI","rawName":"o3-mini","variant":"epoch","setting":"Epoch AI run, accuracy; effort: high","score":96.49,"unit":"%","rank":10,"rankComputedBySite":true,"date":"2025-02-13","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-haiku-4-5","modelName":"Claude Haiku 4.5","provider":"Anthropic","rawName":"Claude Haiku 4.5","variant":"epoch","setting":"Epoch AI run, accuracy","score":96.36,"unit":"%","rank":11,"rankComputedBySite":true,"date":"2025-10-22","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gemini-2-5-pro-may-2025","modelName":"Gemini 2.5 Pro (May 2025)","provider":"Google","rawName":"Gemini 2.5 Pro (May 2025)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Jun 2025","score":95.9,"unit":"%","rank":12,"rankComputedBySite":true,"date":"2025-05-08","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gemini-2-5-pro-mar-2025","modelName":"Gemini 2.5 Pro (Mar 2025)","provider":"Google","rawName":"Gemini 2.5 Pro (Mar 2025)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Mar 2025","score":95.56,"unit":"%","rank":13,"rankComputedBySite":true,"date":"2025-05-07","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-5-nano","modelName":"GPT-5 nano","provider":"OpenAI","rawName":"GPT-5 nano","variant":"epoch","setting":"Epoch AI run, accuracy; effort: medium","score":95.24,"unit":"%","rank":14,"rankComputedBySite":true,"date":"2025-08-20","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"o3-mini","modelName":"o3-mini","provider":"OpenAI","rawName":"o3-mini","variant":"epoch","setting":"Epoch AI run, accuracy; effort: medium","score":95.17,"unit":"%","rank":15,"rankComputedBySite":true,"date":"2025-01-31","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-5-nano","modelName":"GPT-5 nano","provider":"OpenAI","rawName":"GPT-5 nano","variant":"epoch","setting":"Epoch AI run, accuracy; effort: high","score":94.9,"unit":"%","rank":16,"rankComputedBySite":true,"date":"2025-08-20","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"o1","modelName":"o1","provider":"OpenAI","rawName":"o1","variant":"epoch","setting":"Epoch AI run, accuracy; effort: high","score":94.71,"unit":"%","rank":17,"rankComputedBySite":true,"date":"2025-02-13","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"o1","modelName":"o1","provider":"OpenAI","rawName":"o1","variant":"epoch","setting":"Epoch AI run, accuracy; effort: medium","score":94.41,"unit":"%","rank":18,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"deepseek-r1","modelName":"DeepSeek-R1","provider":"DeepSeek","rawName":"DeepSeek-R1","variant":"epoch","setting":"Epoch AI run, accuracy","score":93.05,"unit":"%","rank":19,"rankComputedBySite":true,"date":"2025-01-31","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-3-7-sonnet","modelName":"Claude 3.7 Sonnet","provider":"Anthropic","rawName":"Claude 3.7 Sonnet","variant":"epoch","setting":"Epoch AI run, accuracy; effort: 64k thinking","score":91.16,"unit":"%","rank":20,"rankComputedBySite":true,"date":"2025-03-13","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"grok-3-mini","modelName":"Grok-3 mini","provider":"xAI","rawName":"Grok-3 mini","variant":"epoch","setting":"Epoch AI run, accuracy","score":90.94,"unit":"%","rank":21,"rankComputedBySite":true,"date":"2025-04-10","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-3-7-sonnet","modelName":"Claude 3.7 Sonnet","provider":"Anthropic","rawName":"Claude 3.7 Sonnet","variant":"epoch","setting":"Epoch AI run, accuracy; effort: 32k thinking","score":90.03,"unit":"%","rank":22,"rankComputedBySite":true,"date":"2025-03-12","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"deepseek-r1-distill-llama-70b","modelName":"DeepSeek-R1-Distill-Llama-70B","provider":"DeepSeek","rawName":"DeepSeek-R1-Distill-Llama-70B","variant":"epoch","setting":"Epoch AI run, accuracy","score":89.9,"unit":"%","rank":23,"rankComputedBySite":true,"date":"2025-03-10","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"o1-mini","modelName":"o1-mini","provider":"OpenAI","rawName":"o1-mini","variant":"epoch","setting":"Epoch AI run, accuracy; effort: high","score":89.18,"unit":"%","rank":24,"rankComputedBySite":true,"date":"2025-02-13","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"grok-3","modelName":"Grok 3","provider":"xAI","rawName":"Grok 3","variant":"epoch","setting":"Epoch AI run, accuracy","score":88.75,"unit":"%","rank":25,"rankComputedBySite":true,"date":"2025-04-10","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-4-1-mini","modelName":"GPT-4.1 mini","provider":"OpenAI","rawName":"GPT-4.1 mini","variant":"epoch","setting":"Epoch AI run, accuracy","score":87.29,"unit":"%","rank":26,"rankComputedBySite":true,"date":"2025-04-14","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"deepseek-r1-distill-qwen-14b","modelName":"DeepSeek-R1-Distill-Qwen-14B","provider":"DeepSeek","rawName":"DeepSeek-R1-Distill-Qwen-14B","variant":"epoch","setting":"Epoch AI run, accuracy","score":87.12,"unit":"%","rank":27,"rankComputedBySite":true,"date":"2025-03-11","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-3-7-sonnet","modelName":"Claude 3.7 Sonnet","provider":"Anthropic","rawName":"Claude 3.7 Sonnet","variant":"epoch","setting":"Epoch AI run, accuracy; effort: 16k thinking","score":86.25,"unit":"%","rank":28,"rankComputedBySite":true,"date":"2025-02-26","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-opus-4","modelName":"Claude Opus 4","provider":"Anthropic","rawName":"Claude Opus 4","variant":"epoch","setting":"Epoch AI run, accuracy","score":85.05,"unit":"%","rank":29,"rankComputedBySite":true,"date":"2025-05-22","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-sonnet-4","modelName":"Claude Sonnet 4","provider":"Anthropic","rawName":"Claude Sonnet 4","variant":"epoch","setting":"Epoch AI run, accuracy","score":84.37,"unit":"%","rank":30,"rankComputedBySite":true,"date":"2025-05-22","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"o1-mini","modelName":"o1-mini","provider":"OpenAI","rawName":"o1-mini","variant":"epoch","setting":"Epoch AI run, accuracy; effort: medium","score":84.25,"unit":"%","rank":31,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gemini-2-0-pro","modelName":"Gemini 2.0 Pro","provider":"Google","rawName":"Gemini 2.0 Pro","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Feb 2025","score":83.46,"unit":"%","rank":32,"rankComputedBySite":true,"date":"2025-02-06","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-4-1","modelName":"GPT-4.1","provider":"OpenAI","rawName":"GPT-4.1","variant":"epoch","setting":"Epoch AI run, accuracy","score":83.01,"unit":"%","rank":33,"rankComputedBySite":true,"date":"2025-04-14","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gemini-2-0-flash-feb-2025","modelName":"Gemini 2.0 Flash (Feb 2025)","provider":"Google","rawName":"Gemini 2.0 Flash (Feb 2025)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Feb 2025","score":82.17,"unit":"%","rank":34,"rankComputedBySite":true,"date":"2025-02-06","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"o1-preview","modelName":"o1-preview","provider":"OpenAI","rawName":"o1-preview","variant":"epoch","setting":"Epoch AI run, accuracy","score":81.65,"unit":"%","rank":35,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"mistral-medium-3","modelName":"Mistral Medium 3","provider":"Mistral AI","rawName":"Mistral Medium 3","variant":"epoch","setting":"Epoch AI run, accuracy","score":81.63,"unit":"%","rank":36,"rankComputedBySite":true,"date":"2025-05-07","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-4-5","modelName":"GPT-4.5","provider":"OpenAI","rawName":"GPT-4.5","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Feb 2025","score":78.63,"unit":"%","rank":37,"rankComputedBySite":true,"date":"2025-02-28","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"deepseek-v3-mar-2025","modelName":"DeepSeek-V3 (Mar 2025)","provider":"DeepSeek","rawName":"DeepSeek-V3 (Mar 2025)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Mar 2025","score":75.55,"unit":"%","rank":38,"rankComputedBySite":true,"date":"2025-04-01","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gemma-3-27b","modelName":"Gemma 3 27B","provider":"Google","rawName":"Gemma 3 27B","variant":"epoch","setting":"Epoch AI run, accuracy","score":74.04,"unit":"%","rank":39,"rankComputedBySite":true,"date":"2025-03-13","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"llama-4-maverick","modelName":"Llama 4 Maverick","provider":"Meta","rawName":"Llama 4 Maverick","variant":"epoch","setting":"Epoch AI run, accuracy; effort: FP8","score":73.02,"unit":"%","rank":40,"rankComputedBySite":true,"date":"2025-04-08","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gemini-1-5-pro-sept-2024","modelName":"Gemini 1.5 Pro (Sept 2024)","provider":"Google","rawName":"Gemini 1.5 Pro (Sept 2024)","variant":"epoch","setting":"Epoch AI run, accuracy","score":70.39,"unit":"%","rank":41,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-4-1-nano","modelName":"GPT-4.1 nano","provider":"OpenAI","rawName":"GPT-4.1 nano","variant":"epoch","setting":"Epoch AI run, accuracy","score":70,"unit":"%","rank":42,"rankComputedBySite":true,"date":"2025-04-14","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"qwen3-235b-a22b","modelName":"Qwen3-235B-A22B","provider":"Alibaba (Qwen)","rawName":"Qwen3-235B-A22B","variant":"epoch","setting":"Epoch AI run, accuracy","score":68.86,"unit":"%","rank":43,"rankComputedBySite":true,"date":"2025-06-03","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-3-7-sonnet","modelName":"Claude 3.7 Sonnet","provider":"Anthropic","rawName":"Claude 3.7 Sonnet","variant":"epoch","setting":"Epoch AI run, accuracy","score":68.18,"unit":"%","rank":44,"rankComputedBySite":true,"date":"2025-02-24","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"qwen2-5-max","modelName":"Qwen2.5-Max","provider":"Alibaba (Qwen)","rawName":"Qwen2.5-Max","variant":"epoch","setting":"Epoch AI run, accuracy","score":67.18,"unit":"%","rank":45,"rankComputedBySite":true,"date":"2025-04-01","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"qwen-plus-jan-2025","modelName":"Qwen Plus (Jan 2025)","provider":"Alibaba (Qwen)","rawName":"Qwen Plus (Jan 2025)","variant":"epoch","setting":"Epoch AI run, accuracy","score":65.28,"unit":"%","rank":46,"rankComputedBySite":true,"date":"2025-04-07","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"phi-4","modelName":"Phi-4","provider":"Microsoft Research","rawName":"Phi-4","variant":"epoch","setting":"Epoch AI run, accuracy","score":64.94,"unit":"%","rank":47,"rankComputedBySite":true,"date":"2025-01-31","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"deepseek-v3","modelName":"DeepSeek-V3","provider":"DeepSeek","rawName":"DeepSeek-V3","variant":"epoch","setting":"Epoch AI run, accuracy","score":64.85,"unit":"%","rank":48,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"grok-2-dec-2024","modelName":"Grok-2 (Dec 2024)","provider":"xAI","rawName":"Grok-2 (Dec 2024)","variant":"epoch","setting":"Epoch AI run, accuracy","score":63.52,"unit":"%","rank":49,"rankComputedBySite":true,"date":"2025-02-17","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"qwen2-5-72b","modelName":"Qwen2.5-72B","provider":"Alibaba (Qwen)","rawName":"Qwen2.5-72B","variant":"epoch","setting":"Epoch AI run, accuracy","score":63.17,"unit":"%","rank":50,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"llama-4-scout","modelName":"Llama 4 Scout","provider":"Meta","rawName":"Llama 4 Scout","variant":"epoch","setting":"Epoch AI run, accuracy","score":62.27,"unit":"%","rank":51,"rankComputedBySite":true,"date":"2025-04-08","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gemini-1-5-flash-sep-2024","modelName":"Gemini 1.5 Flash (Sep 2024)","provider":"Google","rawName":"Gemini 1.5 Flash (Sep 2024)","variant":"epoch","setting":"Epoch AI run, accuracy","score":61.87,"unit":"%","rank":52,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-3-5-sonnet-october-2024","modelName":"Claude 3.5 Sonnet (October 2024)","provider":"Anthropic","rawName":"Claude 3.5 Sonnet (October 2024)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Oct 2024","score":56.95,"unit":"%","rank":53,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"qwen-turbo","modelName":"Qwen-Turbo","provider":"Alibaba (Qwen)","rawName":"Qwen-Turbo","variant":"epoch","setting":"Epoch AI run, accuracy","score":56.23,"unit":"%","rank":54,"rankComputedBySite":true,"date":"2025-04-07","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"qwen2-5-32b","modelName":"Qwen2.5-32B","provider":"Alibaba (Qwen)","rawName":"Qwen2.5-32B","variant":"epoch","setting":"Epoch AI run, accuracy","score":56.07,"unit":"%","rank":55,"rankComputedBySite":true,"date":"2025-01-30","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-4o-aug-2024","modelName":"GPT-4o (Aug 2024)","provider":"OpenAI","rawName":"GPT-4o (Aug 2024)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Aug 2024","score":53.28,"unit":"%","rank":56,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-4o-mini","modelName":"GPT-4o mini","provider":"OpenAI","rawName":"GPT-4o mini","variant":"epoch","setting":"Epoch AI run, accuracy","score":52.63,"unit":"%","rank":57,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-3-5-sonnet","modelName":"Claude 3.5 Sonnet","provider":"Anthropic","rawName":"Claude 3.5 Sonnet","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Jun 2024","score":51.68,"unit":"%","rank":58,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-4o-may-2024","modelName":"GPT-4o (May 2024)","provider":"OpenAI","rawName":"GPT-4o (May 2024)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: May 2024","score":51.05,"unit":"%","rank":59,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"mistral-large-2-nov-2024","modelName":"Mistral Large 2 (Nov 2024)","provider":"Mistral AI","rawName":"Mistral Large 2 (Nov 2024)","variant":"epoch","setting":"Epoch AI run, accuracy","score":50.28,"unit":"%","rank":60,"rankComputedBySite":true,"date":"2025-02-25","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-4o-nov-2024","modelName":"GPT-4o (Nov 2024)","provider":"OpenAI","rawName":"GPT-4o (Nov 2024)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Nov 2024","score":49.77,"unit":"%","rank":61,"rankComputedBySite":true,"date":"2025-02-05","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"llama-3-1-405b","modelName":"Llama 3.1-405B","provider":"Meta","rawName":"Llama 3.1-405B","variant":"epoch","setting":"Epoch AI run, accuracy","score":49.77,"unit":"%","rank":61,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"mistral-small-3-1","modelName":"Mistral Small 3.1","provider":"Mistral AI","rawName":"Mistral Small 3.1","variant":"epoch","setting":"Epoch AI run, accuracy","score":46.77,"unit":"%","rank":63,"rankComputedBySite":true,"date":"2025-03-18","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-4-turbo-apr-2024","modelName":"GPT-4 Turbo (Apr 2024)","provider":"OpenAI","rawName":"GPT-4 Turbo (Apr 2024)","variant":"epoch","setting":"Epoch AI run, accuracy","score":46.73,"unit":"%","rank":64,"rankComputedBySite":true,"date":"2025-02-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-3-5-haiku","modelName":"Claude 3.5 Haiku","provider":"Anthropic","rawName":"Claude 3.5 Haiku","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Oct 2024","score":46.36,"unit":"%","rank":65,"rankComputedBySite":true,"date":"2025-03-12","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"mistral-small-3","modelName":"Mistral Small 3","provider":"Mistral AI","rawName":"Mistral Small 3","variant":"epoch","setting":"Epoch AI run, accuracy","score":44.82,"unit":"%","rank":66,"rankComputedBySite":true,"date":"2025-01-30","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"mistral-large-2-jul-2024","modelName":"Mistral Large 2 (Jul 2024)","provider":"Mistral AI","rawName":"Mistral Large 2 (Jul 2024)","variant":"epoch","setting":"Epoch AI run, accuracy","score":44.82,"unit":"%","rank":66,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"tulu-3-t-lu-3-70b","modelName":"Tulu 3 (Tülu 3) 70B","provider":"Ai2","rawName":"Tulu 3 (Tülu 3) 70B","variant":"epoch","setting":"Epoch AI run, accuracy","score":42.66,"unit":"%","rank":68,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"llama-3-3-70b","modelName":"Llama 3.3 70B","provider":"Meta","rawName":"Llama 3.3 70B","variant":"epoch","setting":"Epoch AI run, accuracy","score":41.6,"unit":"%","rank":69,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gemini-1-5-pro-may-2024","modelName":"Gemini 1.5 Pro (May 2024)","provider":"Google","rawName":"Gemini 1.5 Pro (May 2024)","variant":"epoch","setting":"Epoch AI run, accuracy","score":40.75,"unit":"%","rank":70,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-4-turbo-nov-2023","modelName":"GPT-4 Turbo (Nov 2023)","provider":"OpenAI","rawName":"GPT-4 Turbo (Nov 2023)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Nov 2023","score":40.02,"unit":"%","rank":71,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"llama-3-2-90b","modelName":"Llama 3.2 90B","provider":"Meta","rawName":"Llama 3.2 90B","variant":"epoch","setting":"Epoch AI run, accuracy","score":39.44,"unit":"%","rank":72,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"qwen2-72b","modelName":"Qwen2-72B","provider":"Alibaba (Qwen)","rawName":"Qwen2-72B","variant":"epoch","setting":"Epoch AI run, accuracy","score":39.07,"unit":"%","rank":73,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-3-opus","modelName":"Claude 3 Opus","provider":"Anthropic","rawName":"Claude 3 Opus","variant":"epoch","setting":"Epoch AI run, accuracy","score":37.48,"unit":"%","rank":74,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"llama-3-1-70b","modelName":"Llama 3.1-70B","provider":"Meta","rawName":"Llama 3.1-70B","variant":"epoch","setting":"Epoch AI run, accuracy","score":36.68,"unit":"%","rank":75,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-4-turbo-nov-2023","modelName":"GPT-4 Turbo (Nov 2023)","provider":"OpenAI","rawName":"GPT-4 Turbo (Nov 2023)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: January 2024","score":35.41,"unit":"%","rank":76,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gemma-2-27b","modelName":"Gemma 2 27B","provider":"Google","rawName":"Gemma 2 27B","variant":"epoch","setting":"Epoch AI run, accuracy","score":27.89,"unit":"%","rank":77,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"wizardlm-2-8x22b","modelName":"WizardLM-2 8x22B","provider":"Microsoft","rawName":"WizardLM-2 8x22B","variant":"epoch","setting":"Epoch AI run, accuracy","score":25.74,"unit":"%","rank":78,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"yi-1-5-34b","modelName":"Yi-1.5-34B","provider":"01.AI","rawName":"Yi-1.5-34B","variant":"epoch","setting":"Epoch AI run, accuracy","score":25.48,"unit":"%","rank":79,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gemini-1-5-flash-may-2024","modelName":"Gemini 1.5 Flash (May 2024)","provider":"Google","rawName":"Gemini 1.5 Flash (May 2024)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: May 2024","score":25.12,"unit":"%","rank":80,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"mistral-large","modelName":"Mistral Large","provider":"Mistral AI","rawName":"Mistral Large","variant":"epoch","setting":"Epoch AI run, accuracy","score":24.46,"unit":"%","rank":81,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"mixtral-8x22b","modelName":"Mixtral 8x22B","provider":"Mistral AI","rawName":"Mixtral 8x22B","variant":"epoch","setting":"Epoch AI run, accuracy","score":24.24,"unit":"%","rank":82,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-4-jun-2023","modelName":"GPT-4 (Jun 2023)","provider":"OpenAI","rawName":"GPT-4 (Jun 2023)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Jun 2023","score":22.97,"unit":"%","rank":83,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"llama-3-1-8b","modelName":"Llama 3.1-8B","provider":"Meta","rawName":"Llama 3.1-8B","variant":"epoch","setting":"Epoch AI run, accuracy","score":22.88,"unit":"%","rank":84,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"hermes-2-theta-llama-3-70b","modelName":"Hermes 2 Theta Llama-3 70B","provider":"Nous Research","rawName":"Hermes 2 Theta Llama-3 70B","variant":"epoch","setting":"Epoch AI run, accuracy","score":22.69,"unit":"%","rank":85,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"llama-3-70b","modelName":"Llama 3-70B","provider":"Meta","rawName":"Llama 3-70B","variant":"epoch","setting":"Epoch AI run, accuracy","score":22.55,"unit":"%","rank":86,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gemma-2-9b","modelName":"Gemma 2 9B","provider":"Google","rawName":"Gemma 2 9B","variant":"epoch","setting":"Epoch AI run, accuracy","score":21.01,"unit":"%","rank":87,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-3-sonnet","modelName":"Claude 3 Sonnet","provider":"Anthropic","rawName":"Claude 3 Sonnet","variant":"epoch","setting":"Epoch AI run, accuracy","score":18.17,"unit":"%","rank":88,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"phi-3-medium-14b","modelName":"phi-3-medium 14B","provider":"Microsoft","rawName":"phi-3-medium 14B","variant":"epoch","setting":"Epoch AI run, accuracy","score":17.56,"unit":"%","rank":89,"rankComputedBySite":true,"date":"2025-01-31","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-3-5-turbo-nov-2023","modelName":"GPT-3.5 Turbo (Nov 2023)","provider":"OpenAI","rawName":"GPT-3.5 Turbo (Nov 2023)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Nov 2023","score":15.89,"unit":"%","rank":90,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"ministral-8b","modelName":"Ministral 8B","provider":"Mistral AI","rawName":"Ministral 8B","variant":"epoch","setting":"Epoch AI run, accuracy","score":14.94,"unit":"%","rank":91,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-3-haiku","modelName":"Claude 3 Haiku","provider":"Anthropic","rawName":"Claude 3 Haiku","variant":"epoch","setting":"Epoch AI run, accuracy","score":14.88,"unit":"%","rank":92,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"ministral-3b","modelName":"Ministral 3B","provider":"Mistral AI","rawName":"Ministral 3B","variant":"epoch","setting":"Epoch AI run, accuracy","score":14.44,"unit":"%","rank":93,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"claude-2","modelName":"Claude 2","provider":"Anthropic","rawName":"Claude 2","variant":"epoch","setting":"Epoch AI run, accuracy","score":11.73,"unit":"%","rank":94,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"dbrx","modelName":"DBRX","provider":"Databricks","rawName":"DBRX","variant":"epoch","setting":"Epoch AI run, accuracy","score":11.65,"unit":"%","rank":95,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gpt-3-5-turbo-jan-2024","modelName":"GPT-3.5 Turbo (Jan 2024)","provider":"OpenAI","rawName":"GPT-3.5 Turbo (Jan 2024)","variant":"epoch","setting":"Epoch AI run, accuracy; effort: Jan 2024","score":11.63,"unit":"%","rank":96,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"gemini-1-0-pro","modelName":"Gemini 1.0 Pro","provider":"Google","rawName":"Gemini 1.0 Pro","variant":"epoch","setting":"Epoch AI run, accuracy","score":11.24,"unit":"%","rank":97,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"mistral-nemo","modelName":"Mistral NeMo","provider":"Mistral AI","rawName":"Mistral NeMo","variant":"epoch","setting":"Epoch AI run, accuracy","score":10.83,"unit":"%","rank":98,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"mixtral-8x7b","modelName":"Mixtral 8x7B","provider":"Mistral AI","rawName":"Mixtral 8x7B","variant":"epoch","setting":"Epoch AI run, accuracy","score":9.29,"unit":"%","rank":99,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"deepseek-llm-67b","modelName":"DeepSeek LLM 67B","provider":"DeepSeek","rawName":"DeepSeek LLM 67B","variant":"epoch","setting":"Epoch AI run, accuracy","score":6.39,"unit":"%","rank":100,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"llama-3-8b","modelName":"Llama 3-8B","provider":"Meta","rawName":"Llama 3-8B","variant":"epoch","setting":"Epoch AI run, accuracy","score":6.13,"unit":"%","rank":101,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"yi-34b","modelName":"Yi-34B","provider":"01.AI","rawName":"Yi-34B","variant":"epoch","setting":"Epoch AI run, accuracy","score":5.15,"unit":"%","rank":102,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"mistral-7b-v0-3","modelName":"Mistral 7B v0.3","provider":"Mistral AI","rawName":"Mistral 7B v0.3","variant":"epoch","setting":"Epoch AI run, accuracy","score":3.6,"unit":"%","rank":103,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"},{"model":"llama-2-70b","modelName":"Llama 2-70B","provider":"Meta","rawName":"Llama 2-70B","variant":"epoch","setting":"Epoch AI run, accuracy","score":3.29,"unit":"%","rank":104,"rankComputedBySite":true,"date":"2025-01-27","sourceType":"independent","source":"epoch","sourceUrl":"https://epoch.ai/benchmarks","collected":"2026-10-10"}]}