⌕Is ___ SOTA?
benchmarks / gpqa-diamond

GPQA Diamond

accuracy. Status display_only. Contamination risk med. Maintainer Epoch AI (external items). Source.

Display only.This benchmark cannot decide a crown under methodology v1.0.
ModelScoreDateKind
DeepSeek V4.1 Flash0.89835858585858592026-09-09observation
Claude Haiku 5.50.89646464646464652026-10-07observation
GPT-6.1 Sol0.95391414141414142026-09-29observation
Claude Sonnet 5.50.95580808080808082026-09-28observation
GPT-6 Luna0.90467171717171712026-09-22observation
Grok 4.70.927398989898992026-09-21observation
GPT-6 Sol0.94318181818181822026-09-22observation
Claude Opus 5.50.90593434343434342026-09-22observation
Qwen3 8 Max 09020.9229797979797982026-09-01observation
Gemini 3.8 Flash0.95391414141414142026-09-02observation
Granite 4.0 Micro0.28282828282828282025-10-02observation
Seed Oss 36B Instruct0.71527777777777782025-08-20observation
Qwen3 32B0.54103535353535352025-04-29observation
Qwen3 30B A3b Thinking 25070.70075757575757582025-07-30observation
Qwen3 30B A3b Instruct 25070.55618686868686872025-07-29observation
Qwen3 30B A3b0.50378787878787882025-04-29observation
Qwen2 5 7B Instruct0.35479797979797982024-09-19observation
Mistral Small 3.1 25030.419191919191919172025-03-17observation
Mistral Small 3 25010.472853535353535362025-01-30observation
Magistral Small 25090.4760101010101012025-09-18observation
Llama 3.2 1B Instruct0.239267676767676772024-09-24observation
GPT-6 Astra0.95770202020202022026-09-03observation
Granite 4.0 1B0.23989898989898992025-10-28observation
Qwen3 5 35B A3b0.81186868686868692026-02-24observation
Qwen3 4B Instruct 25070.45833333333333332025-08-06observation
Qwen3 4B0.43560606060606062025-04-29observation
Qwen3 14B0.53409090909090912025-04-29observation
Qwen3 1 7B0.30618686868686872025-04-29observation
Mistral Small 3.2 25060.490530303030303042025-06-20observation
Granite 4.0 350m0.112373737373737382025-10-28observation