supermarket

Benchmark

GPQA Diamond

Measures: The hardest subset of GPQA's graduate-level, 'Google-proof' multiple-choice questions in biology, physics and chemistry written by domain experts; scored as accuracy (random guess 25%).

Maintainer
GPQA authors (Rein et al.); no official leaderboard. Independent runs by Epoch AI and Artificial Analysis
Introduced
2023
Status
Active
Saturation
Saturated
Typical provenance
Mixed
In scoring set
No

Paper

Leaderboard

Show independent results only

Every result we hold, whoever ran it. Only independent and third-party rows count towards a score.

Leaderboard
ModelAccuracyProvenanceSourceEvaluated
GPT-6 Astra
OpenAI · Max effort
95.8%IndependentEpoch AI30 Aug 2026
Gemini 3.8 Flash
Google DeepMind · High effort
95.4%IndependentEpoch AI2 Sep 2026
Gemini 3.7 Flash
Google DeepMind · High effort
94.8%IndependentEpoch AI14 Aug 2026
GPT-5.4 Pro
OpenAI · Xhigh effort
94.6%IndependentEpoch AI20 Mar 2026
Gemini 3.1 Pro
Google DeepMind · High effort
94.4%IndependentEpoch AI6 Aug 2026
Gemini 3.6 Flash
Google DeepMind · High effort
94.1%IndependentEpoch AI2 Aug 2026
Gemini 3.1 Pro
Google DeepMind
94.1%IndependentEpoch AI20 Feb 2026
Grok 4.6
SpaceXAI (xAI) · High effort
94.0%IndependentEpoch AI12 Aug 2026
Claude Opus 5
Anthropic · Max effort
93.9%IndependentEpoch AI24 Jul 2026
Kimi K3
Moonshot AI
93.5%Self-reportedModel Card via Hugging Face Community Evals—
GPT-5.6 Sol
OpenAI · Max effort
93.5%IndependentEpoch AI9 Jul 2026
Grok 4.5
SpaceXAI (xAI) · High effort
93.4%IndependentEpoch AI8 Jul 2026
GPT-5.6 Terra
OpenAI · Max effort
93.3%IndependentEpoch AI9 Jul 2026
GPT-5.4
OpenAI · Xhigh effort
93.3%IndependentEpoch AI6 Mar 2026
Grok 4.6
SpaceXAI (xAI) · Xhigh effort
93.2%IndependentEpoch AI14 Aug 2026
Kimi K3
Moonshot AI · Max effort
93.1%IndependentEpoch AI16 Jul 2026
Claude Opus 5
Anthropic
92.9%IndependentEpoch AI6 Aug 2026
Gemini 3.5 Flash
Google DeepMind · High effort
92.8%IndependentEpoch AI22 May 2026
Qwen3.8-Max
Alibaba Group · Xhigh effort
92.7%IndependentEpoch AI4 Aug 2026
Gemini 3 Pro
Google DeepMind
92.6%IndependentEpoch AI19 Nov 2025
Qwen3.8-2.4T-A95B
Alibaba Group
92.6%Self-reportedModel Card via Hugging Face Community Evals—
Hy4 preview
Tencent
92.3%Self-reportedModel Card via Hugging Face Community Evals—
Qwen3.8-Max
Alibaba Group · Xhigh effort
92.3%IndependentEpoch AI2 Sep 2026
Kimi K3
Moonshot AI · High effort
91.9%IndependentEpoch AI7 Aug 2026
GLM-5.2
Z.ai (Zhipu AI) · Max effort
91.9%IndependentEpoch AI24 Jun 2026
Qwen3.8-Flash-Next
Alibaba Group
91.7%Self-reportedQwen3.8-Flash-Next model card via Hugging Face Community Evals—
DeepSeek-V4-Pro
DeepSeek · Max effort
91.7%IndependentEpoch AI18 Aug 2026
GPT-5.6 Luna
OpenAI · Max effort
91.6%IndependentEpoch AI9 Jul 2026
GPT-5.2
OpenAI · Xhigh effort
91.4%IndependentEpoch AI13 Dec 2025
GLM-5.2
Z.ai (Zhipu AI)
91.2%Self-reportedModel Card via Hugging Face Community Evals—
Claude Opus 4.8
Anthropic · Max effort
91.0%IndependentEpoch AI7 Jun 2026
DeepSeek-V4-Flash
DeepSeek · Max effort
91.0%IndependentEpoch AI2 Aug 2026
DeepSeek-V4-Pro
DeepSeek · High effort
90.9%IndependentEpoch AI6 Aug 2026
GLM-5.3
Z.ai (Zhipu AI) · Max effort
90.9%IndependentEpoch AI24 Aug 2026
MiniMax-M3
MiniMax
90.9%IndependentEpoch AI10 Aug 2026
Qwen3.7-Max
Alibaba Group
90.9%IndependentEpoch AI7 Aug 2026
DeepSeek-V4.1-Flash
DeepSeek
90.9%Self-reportedDeepSeek-V4.1-Flash model card via Hugging Face Community Evals—
Kimi K2.6
Moonshot AI
90.8%IndependentEpoch AI1 May 2026
GPT-5.5
OpenAI · Low effort
90.7%IndependentEpoch AI5 May 2026
Claude Opus 5.5
Anthropic · Max effort
90.6%IndependentEpoch AI22 Sep 2026
Claude Opus 4.6
Anthropic · 32K effort
90.5%IndependentEpoch AI6 Feb 2026
Claude Sonnet 5
Anthropic · Xhigh effort
90.5%IndependentEpoch AI1 Jul 2026
Kimi K2.6
Moonshot AI
90.5%Self-reportedModel Card via Hugging Face Community Evals—
Hy3
Tencent
90.4%Self-reportedModel Card via Hugging Face Community Evals—
Claude Opus 4.7
Anthropic · Xhigh effort
90.2%IndependentEpoch AI17 Apr 2026
GLM-5.3-Flash
Z.ai (Zhipu AI) · Max effort
90.2%IndependentEpoch AI26 Aug 2026
GLM-5.1
Z.ai (Zhipu AI)
89.9%IndependentEpoch AI10 Aug 2026
GPT-5.4
OpenAI · High effort
89.9%IndependentEpoch AI15 Jul 2026
GPT-5.6 Sol
OpenAI · Low effort
89.9%IndependentEpoch AI7 Aug 2026
DeepSeek-V4-Pro
DeepSeek · Max effort
89.6%IndependentEpoch AI16 Jun 2026
Inkling Small
Thinking Machines Lab
89.5%Self-reportedModel Card via Hugging Face Community Evals—
Gemini 3 Flash
Google DeepMind · High effort
89.4%IndependentEpoch AI6 Aug 2026
Grok 4.20
SpaceXAI (xAI)
89.3%IndependentEpoch AI13 Jul 2026
Qwen3.8-27B
Alibaba Group
89.2%Self-reportedQwen3.8-27B model card via Hugging Face Community Evals—
Gemini 3.5 Flash
Google DeepMind · Low effort
88.9%IndependentEpoch AI6 Aug 2026
GPT-5.4
OpenAI · Medium effort
88.9%IndependentEpoch AI15 Jul 2026
Grok 4.3
SpaceXAI (xAI) · High effort
88.8%IndependentEpoch AI17 Jun 2026
Claude Opus 4.6
Anthropic · 64K effort
88.8%IndependentEpoch AI6 Feb 2026
Inkling Small
Thinking Machines Lab · Xhigh effort
88.5%IndependentEpoch AI14 Aug 2026
Qwen3.5-397B-A17B
Alibaba Group
88.4%Self-reportedModel Card via Hugging Face Community Evals—
Claude Opus 4.6
Anthropic · Max effort
88.4%IndependentEpoch AI6 Aug 2026
Claude Opus 4.8
Anthropic · Low effort
88.4%IndependentEpoch AI6 Aug 2026
Qwen3.6-Plus
Alibaba Group
88.4%IndependentEpoch AI7 Aug 2026
Ring-2.6-1T
Ant Group
88.3%Self-reportedModel Card via Hugging Face Community Evals—
Inkling
Thinking Machines Lab · Xhigh effort
88.3%IndependentEpoch AI5 Aug 2026
GPT-5.2
OpenAI · High effort
88.2%IndependentEpoch AI11 Dec 2025
Claude Opus 5
Anthropic · Low effort
87.9%IndependentEpoch AI6 Aug 2026
GLM-5.2
Z.ai (Zhipu AI) · Low effort
87.9%IndependentEpoch AI10 Aug 2026
GPT-5.2
OpenAI · Medium effort
87.9%IndependentEpoch AI11 Dec 2025
Kimi K2.7 Code
Moonshot AI
87.9%IndependentEpoch AI7 Aug 2026
Qwen3.7-Plus
Alibaba Group
87.9%IndependentEpoch AI7 Aug 2026
GLM-5
Z.ai (Zhipu AI)
87.8%IndependentEpoch AI12 Feb 2026
Qwen3.6-27B
Alibaba Group
87.8%Self-reportedModel Card via Hugging Face Community Evals—
GPT-5.1
OpenAI · High effort
87.6%IndependentEpoch AI13 Nov 2025
Kimi K2.5
Moonshot AI
87.6%Self-reportedModel Card via Hugging Face Community Evals—
Claude Sonnet 4.6
Anthropic · 32K effort
87.4%IndependentEpoch AI20 Feb 2026
GPT-5.6 Terra
OpenAI · Low effort
87.4%IndependentEpoch AI7 Aug 2026
Qwen3.6-Max-Preview
Alibaba Group
87.4%IndependentEpoch AI7 Aug 2026
Hy3 preview
Tencent
87.2%Self-reportedModel Card via Hugging Face Community Evals—
Inkling
Thinking Machines Lab
87.2%Self-reportedModel Card via Hugging Face Community Evals—
Grok 4
SpaceXAI (xAI)
87.0%IndependentEpoch AI—
Nemotron 3 Ultra 550B-A55B
NVIDIA
87.0%Self-reportedModel Card via Hugging Face Community Evals—
GPT-5.4 mini
OpenAI · Xhigh effort
86.9%IndependentEpoch AI7 Aug 2026
Qwen3.5-122B-A10B
Alibaba Group
86.6%Self-reportedModel Card via Hugging Face Community Evals—
Claude Opus 4.7
Anthropic · Max effort
86.4%IndependentEpoch AI6 Aug 2026
Gemini 3.5 Flash
Google DeepMind · Minimal effort
86.4%IndependentEpoch AI15 Jul 2026
Gemini 3.6 Flash
Google DeepMind · Low effort
86.4%IndependentEpoch AI6 Aug 2026
Qwen3.5-397B-A17B
Alibaba Group · None effort
86.4%IndependentEpoch AI7 Aug 2026
Solar Open 2 (250B)
Upstage
86.3%Self-reportedModel Card via Hugging Face Community Evals—
GLM-5.1
Z.ai (Zhipu AI)
86.2%Self-reportedModel Card via Hugging Face Community Evals—
GPT-5
OpenAI · High effort
86.2%IndependentEpoch AI29 Oct 2025
Claude Opus 4.5
Anthropic · 32K effort
86.0%IndependentEpoch AI24 Nov 2025
GLM-5
Z.ai (Zhipu AI)
86.0%Self-reportedModel Card via Hugging Face Community Evals—
Qwen3.6-35B-A3B
Alibaba Group
86.0%Self-reportedModel Card via Hugging Face Community Evals—
Claude Fable 5
Anthropic · Max effort
85.9%IndependentEpoch AI6 Aug 2026
Gemini 3.6 Flash
Google DeepMind · Minimal effort
85.9%IndependentEpoch AI7 Aug 2026
Qwen3.5-397B-A17B
Alibaba Group
85.9%IndependentEpoch AI7 Aug 2026
Qwen3.6-27B
Alibaba Group
85.9%IndependentEpoch AI7 Aug 2026
GLM-4.7
Z.ai (Zhipu AI)
85.7%Self-reportedModel Card via Hugging Face Community Evals—
A.X K2
SK Telecom
85.6%Self-reportedModel Card via Hugging Face Community Evals—
Qwen3.5-27B
Alibaba Group
85.5%Self-reportedModel Card via Hugging Face Community Evals—
Claude Opus 4.5
Anthropic · 16K effort
85.5%IndependentEpoch AI25 Nov 2025
Claude Opus 4.8
Anthropic · None effort
85.4%IndependentEpoch AI6 Aug 2026
GPT-5
OpenAI · Medium effort
85.4%IndependentEpoch AI7 Aug 2025
Gemini 2.5 Pro
Google DeepMind
85.3%IndependentEpoch AI16 Nov 2025
MiniMax-M2.5
MiniMax
85.2%Self-reportedModel Card via Hugging Face Community Evals—
GPT-5.1
OpenAI · Medium effort
85.0%IndependentEpoch AI17 Nov 2025
Gemini 2.5 Pro
Google DeepMind
84.8%IndependentEpoch AI5 Jun 2025
GPT-5.4
OpenAI · Low effort
84.8%IndependentEpoch AI15 Jul 2026
Kimi K3
Moonshot AI · Low effort
84.8%IndependentEpoch AI7 Aug 2026
Qwen3.5-Plus
Alibaba Group
84.8%IndependentEpoch AI7 Aug 2026
Qwen3.6-27B
Alibaba Group · None effort
84.8%IndependentEpoch AI7 Aug 2026
Qwen3.6-35B-A3B
Alibaba Group · None effort
84.8%IndependentEpoch AI7 Aug 2026
Kimi K2 Thinking
Moonshot AI · Thinking effort
84.5%Self-reportedModel Card via Hugging Face Community Evals—
Kimi K2 Thinking
Moonshot AI
84.2%IndependentEpoch AI11 Nov 2025
Gemini 2.5 Pro
Google DeepMind
83.8%IndependentEpoch AI31 Mar 2025
Qwen3.6-35B-A3B
Alibaba Group
83.8%IndependentEpoch AI7 Aug 2026
GPT-5.4 mini
OpenAI · High effort
83.6%IndependentEpoch AI15 Apr 2026
Muse Glimmer 30B
Meta
83.5%Self-reportedMuse Glimmer-30B model card via Hugging Face Community Evals—
Step 3.5 Flash
StepFun
83.5%Self-reportedStep 3.5 Flash Paper via Hugging Face Community Evals—
Qwen3.5-35B-A3B
Alibaba Group
83.5%IndependentEpoch AI28 Aug 2026
DeepSeek-V3.2
DeepSeek
83.4%IndependentEpoch AI16 Dec 2025
Claude Fable 5
Anthropic · High effort
83.3%IndependentEpoch AI6 Aug 2026
Claude Sonnet 4.6
Anthropic · High effort
83.3%IndependentEpoch AI13 Jul 2026
Claude Sonnet 4.6
Anthropic · Medium effort
83.3%IndependentEpoch AI13 Jul 2026
Gemini 3.5 Flash-Lite
Google DeepMind · High effort
83.3%IndependentEpoch AI6 Aug 2026
GLM-4.7
Z.ai (Zhipu AI)
83.3%IndependentEpoch AI29 Jan 2026
MiMo-V2-Flash
Xiaomi
83.3%Self-reportedEvalEval via Hugging Face Community Evals—
Qwen3.6-Flash
Alibaba Group
83.3%IndependentEpoch AI7 Aug 2026
Gemini 3 Flash
Google DeepMind
83.2%IndependentEpoch AI17 Dec 2025
GPT-5.6 Sol
OpenAI · None effort
82.8%IndependentEpoch AI7 Aug 2026
GPT-5.2
OpenAI · Low effort
82.7%IndependentEpoch AI11 Dec 2025
Nemotron 3 Super 120B-A12B
NVIDIA
82.7%Self-reportedModel Card via Hugging Face Community Evals—
Claude Sonnet 4.5
Anthropic · 59K effort
82.3%IndependentEpoch AI28 Oct 2025
GPT-5.6 Luna
OpenAI · Low effort
82.3%IndependentEpoch AI7 Aug 2026
Qwen3.5-Flash
Alibaba Group
82.3%IndependentEpoch AI7 Aug 2026
Qwen3.7-Flash
Alibaba Group
82.3%IndependentEpoch AI7 Aug 2026
Gemini 3.1 Flash-Lite
Google DeepMind · High effort
81.8%IndependentEpoch AI6 Aug 2026
o3
OpenAI · High effort
81.8%IndependentEpoch AI16 Apr 2025
Qwen3.7-Plus
Alibaba Group · None effort
81.8%IndependentEpoch AI7 Aug 2026
Claude Sonnet 4.5
Anthropic · 32K effort
81.7%IndependentEpoch AI21 Oct 2025
MiniMax-M3
MiniMax · None effort
81.3%IndependentEpoch AI10 Aug 2026
Qwen3.5-35B-A3B
Alibaba Group · None effort
81.2%IndependentEpoch AI28 Aug 2026
gpt-oss-120b
OpenAI
80.9%Self-reportedGPT-OSS Model Card via Hugging Face Community Evals—
MiniMax-M2.1
MiniMax
80.8%Self-reportedEvalEval via Hugging Face Community Evals—
o3
OpenAI · Medium effort
80.8%IndependentEpoch AI7 Aug 2026
Qwen3.7-Flash
Alibaba Group · None effort
80.8%IndependentEpoch AI7 Aug 2026
Claude Opus 4.5
Anthropic
80.7%IndependentEpoch AI24 Nov 2025
EXAONE-4.5-33B
LG AI Research
80.5%Self-reportedModel Card via Hugging Face Community Evals—
LongCat-Flash-Thinking
Meituan
80.5%Self-reportedModel Card via Hugging Face Community Evals—
Claude Sonnet 5
Anthropic · Max effort
80.3%IndependentEpoch AI6 Aug 2026
Qwen3-235B-A22B-Thinking-2507
Alibaba Group
80.1%IndependentEpoch AI11 Dec 2025
o3
OpenAI · Low effort
79.8%IndependentEpoch AI15 Jul 2026
Claude 3.7 Sonnet
Anthropic · 64K effort
79.7%IndependentEpoch AI26 May 2025
o4-mini
OpenAI · High effort
79.6%IndependentEpoch AI16 Apr 2025
Claude Sonnet 4
Anthropic · 59K effort
79.2%IndependentEpoch AI26 May 2025
K-EXAONE-236B-A23B
LG AI Research
79.1%Self-reportedModel Card via Hugging Face Community Evals—
Qwen3.5-9B
Alibaba Group
79.0%IndependentEpoch AI27 Aug 2026
Qwen3.5-9B
Alibaba Group · None effort
78.9%IndependentEpoch AI27 Aug 2026
Claude Fable 5
Anthropic · Low effort
78.8%IndependentEpoch AI6 Aug 2026
Claude Sonnet 4.5
Anthropic · 16K effort
78.8%IndependentEpoch AI28 Oct 2025
Claude Sonnet 4.6
Anthropic · Max effort
78.8%IndependentEpoch AI6 Aug 2026
GPT-5.4 nano
OpenAI · High effort
78.5%IndependentEpoch AI14 Apr 2026
Claude Sonnet 4
Anthropic · 32K effort
78.3%IndependentEpoch AI22 May 2025
o4-mini
OpenAI · Medium effort
77.8%IndependentEpoch AI7 Aug 2026
Claude Opus 4.1
Anthropic · 16K effort
77.3%IndependentEpoch AI5 Aug 2025
GPT-5.5
OpenAI · None effort
77.3%IndependentEpoch AI7 Aug 2026
GPT-5.6 Terra
OpenAI · None effort
77.3%IndependentEpoch AI7 Aug 2026
o3-mini
OpenAI · High effort
77.0%IndependentEpoch AI13 Feb 2025
Claude 3.7 Sonnet
Anthropic · 32K effort
76.8%IndependentEpoch AI10 Mar 2025
Claude 3.7 Sonnet
Anthropic · 16K effort
76.8%IndependentEpoch AI26 Feb 2025
Claude Opus 4.1
Anthropic · 27K effort
76.8%IndependentEpoch AI5 Aug 2025
o1
OpenAI · High effort
76.8%IndependentEpoch AI13 Feb 2025
DeepSeek-R1
DeepSeek
76.3%IndependentEpoch AI29 May 2025
Claude Opus 4
Anthropic · 16K effort
76.3%IndependentEpoch AI22 May 2025
Grok 3 Mini
SpaceXAI (xAI) · Low effort
76.3%IndependentEpoch AI10 Apr 2025
Claude Sonnet 4
Anthropic · 16K effort
75.8%IndependentEpoch AI22 May 2025
Gemini 3.5 Flash-Lite
Google DeepMind · Low effort
75.8%IndependentEpoch AI6 Aug 2026
Gemma 4 31B
Google DeepMind · Minimal effort
75.8%IndependentEpoch AI6 Aug 2026
gpt-oss-120b
OpenAI · High effort
75.8%IndependentEpoch AI11 Dec 2025
Grok 3
SpaceXAI (xAI)
75.8%IndependentEpoch AI26 May 2025
o1
OpenAI · Medium effort
75.8%IndependentEpoch AI27 Jan 2025
Grok 3 Mini
SpaceXAI (xAI) · High effort
75.5%IndependentEpoch AI26 May 2025
Nemotron 3.5 Lightning 30B-A3B
NVIDIA
75.4%Self-reportedNVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 model card via Hugging Face Community Evals—
o4-mini
OpenAI · Low effort
75.3%IndependentEpoch AI11 Jul 2026
GLM-4.7-Flash
Z.ai (Zhipu AI)
75.2%Self-reportedModel Card via Hugging Face Community Evals—
GPT-5 mini
OpenAI · High effort
75.0%IndependentEpoch AI30 Oct 2025
GPT-5.4
OpenAI · None effort
74.7%IndependentEpoch AI15 Jul 2026
o3-mini
OpenAI · Medium effort
74.3%IndependentEpoch AI31 Jan 2025
Gemini 3.1 Flash-Lite
Google DeepMind · Low effort
74.2%IndependentEpoch AI6 Aug 2026
Gemini 3.5 Flash-Lite
Google DeepMind · Minimal effort
74.2%IndependentEpoch AI6 Aug 2026
o1
OpenAI · Low effort
74.2%IndependentEpoch AI20 Jul 2026
gpt-oss-20b
OpenAI
74.2%Self-reportedGPT-OSS Model Card via Hugging Face Community Evals—
A.X K1
SK Telecom
74.0%Self-reportedA.X K1 model card via Hugging Face Community Evals—
Claude Sonnet 4.5
Anthropic
73.7%IndependentEpoch AI29 Sep 2025
Gemini 3.1 Flash-Lite
Google DeepMind · Minimal effort
73.7%IndependentEpoch AI6 Aug 2026
Claude Opus 4.1
Anthropic
73.2%IndependentEpoch AI5 Aug 2025
DeepSeek-V4-Pro
DeepSeek · None effort
73.2%IndependentEpoch AI6 Aug 2026
Gemma 4 26B A4B
Google DeepMind · Minimal effort
73.2%IndependentEpoch AI6 Aug 2026
GPT-5.2
OpenAI · None effort
73.2%IndependentEpoch AI13 Jul 2026
Qwen3-Max
Alibaba Group
72.6%IndependentEpoch AI6 Oct 2025
GPT-5.4 nano
OpenAI · Low effort
72.2%IndependentEpoch AI7 Aug 2026
Qwen3-30B-A3B-Thinking-2507
Alibaba Group
72.2%Self-reportedEvalEval via Hugging Face Community Evals—
DeepSeek-R1
DeepSeek
71.7%IndependentEpoch AI26 May 2025
GLM-4.5-Air
Z.ai (Zhipu AI)
71.7%Self-reportedEvalEval via Hugging Face Community Evals—
GPT-5
OpenAI · Minimal effort
71.7%IndependentEpoch AI20 Jul 2026
GPT-5 mini
OpenAI · Minimal effort
71.7%IndependentEpoch AI7 Aug 2026
GPT-5 mini
OpenAI · Medium effort
71.7%IndependentEpoch AI7 Aug 2025
Seed-OSS-36B-Instruct
ByteDance
71.5%IndependentEpoch AI30 Aug 2026
DeepSeek-R1
DeepSeek
71.5%Self-reportedModel Card via Hugging Face Community Evals—
Claude Haiku 4.5
Anthropic · 32K effort
71.2%IndependentEpoch AI22 Oct 2025
DeepSeek-V3.2
DeepSeek
71.2%IndependentEpoch AI16 Jul 2026
GLM-5.2
Z.ai (Zhipu AI) · None effort
71.2%IndependentEpoch AI10 Aug 2026
Qwen3-235B-A22B
Alibaba Group
70.7%IndependentEpoch AI3 Jun 2025
Qwen3-30B-A3B-Thinking-2507
Alibaba Group
70.1%IndependentEpoch AI30 Aug 2026
GPT-5 nano
OpenAI · High effort
69.4%IndependentEpoch AI30 Oct 2025
Claude Opus 4
Anthropic
69.2%IndependentEpoch AI22 May 2025
GPT-4.5
OpenAI
68.7%IndependentEpoch AI28 Feb 2025
o3-mini
OpenAI · Low effort
68.2%IndependentEpoch AI20 Jul 2026
GPT-5 nano
OpenAI · Medium effort
67.4%IndependentEpoch AI7 Aug 2025
Llama 4 Maverick
Meta
67.0%IndependentEpoch AI8 Apr 2025
GPT-4.1
OpenAI
66.9%IndependentEpoch AI14 Apr 2025
MiMo-V2.5-Pro
Xiaomi
66.7%Self-reportedModel Card via Hugging Face Community Evals—
Claude Sonnet 4
Anthropic
66.7%IndependentEpoch AI22 May 2025
Gemini 2.5 Pro
Google DeepMind
66.7%IndependentEpoch AI3 Jun 2025
GPT-5.1
OpenAI · None effort
66.7%IndependentEpoch AI7 Aug 2026
Claude 3.7 Sonnet
Anthropic
66.0%IndependentEpoch AI24 Feb 2025
GPT-4.1 mini
OpenAI
65.8%IndependentEpoch AI14 Apr 2025
Qwen3-4B-Thinking-2507
Alibaba Group
65.8%Self-reportedModel Card via Hugging Face Community Evals—
Qwen3-32B
Alibaba Group
65.7%IndependentEpoch AI28 Aug 2026
Gemini 2.0 Pro
Google DeepMind
65.7%IndependentEpoch AI7 Feb 2025
QwQ-Plus
Alibaba Group
65.4%IndependentEpoch AI11 Apr 2025
QwQ-32B
Alibaba Group
65.3%IndependentEpoch AI28 Aug 2026
DeepSeek-R1-Distill-Qwen-32B
DeepSeek
64.1%IndependentEpoch AI28 Aug 2026
Gemini 2.0 Flash
Google DeepMind
64.1%IndependentEpoch AI6 Feb 2025
GPT-5.4 mini
OpenAI · None effort
64.1%IndependentEpoch AI7 Aug 2026
Qwen3-14B
Alibaba Group
63.8%IndependentEpoch AI28 Aug 2026
GPT-5.6 Luna
OpenAI · None effort
63.6%IndependentEpoch AI7 Aug 2026
o1-mini
OpenAI · High effort
62.4%IndependentEpoch AI13 Feb 2025
Qwen3-4B-Instruct-2507
Alibaba Group
62.0%Self-reportedModel Card via Hugging Face Community Evals—
Qwen3-30B-A3B
Alibaba Group
61.7%IndependentEpoch AI30 Aug 2026
gpt-oss-20b
OpenAI · Medium effort
60.8%IndependentEpoch AI27 Aug 2026
GLM-4.7-Flash
Z.ai (Zhipu AI)
60.5%IndependentEpoch AI30 Aug 2026
Claude Haiku 4.5
Anthropic
60.5%IndependentEpoch AI16 Oct 2025
Mistral Medium 3
Mistral AI
59.5%IndependentEpoch AI7 May 2025
o1-mini
OpenAI · Medium effort
59.5%IndependentEpoch AI27 Jan 2025
DeepSeek-V3
DeepSeek
58.2%Self-reportedEvalEval via Hugging Face Community Evals—
GPT-5 nano
OpenAI · Low effort
57.6%IndependentEpoch AI13 Jul 2026
Gemini 1.5 Pro
Google DeepMind
57.2%IndependentEpoch AI27 Jan 2025
Gemini 2.0 Flash Thinking
Google DeepMind
57.1%IndependentEpoch AI6 Feb 2025
Qwen3-8B
Alibaba Group
56.8%IndependentEpoch AI27 Aug 2026
DeepSeek-V3
DeepSeek
56.5%IndependentEpoch AI27 Jan 2025
Qwen2.5-Max
Alibaba Group
56.1%IndependentEpoch AI1 Apr 2025
Magistral Small 1.0
Mistral AI
56.1%IndependentEpoch AI6 Aug 2026
Phi-4
Microsoft
56.1%IndependentEpoch AI31 Jan 2025
DeepSeek-R1-Distill-Llama-70B
DeepSeek
55.7%IndependentEpoch AI10 Mar 2025
Qwen3-30B-A3B-Instruct-2507
Alibaba Group
55.6%IndependentEpoch AI30 Aug 2026
GPT-5.4 nano
OpenAI · None effort
55.6%IndependentEpoch AI7 Aug 2026
Claude 3.5 Sonnet
Anthropic
55.3%IndependentEpoch AI27 Jan 2025
Qwen3-32B
Alibaba Group · None effort
54.1%IndependentEpoch AI30 Aug 2026
Claude 3.5 Sonnet
Anthropic
54.0%IndependentEpoch AI27 Jan 2025
Grok 2
SpaceXAI (xAI)
53.8%IndependentEpoch AI27 Jan 2025
Qwen3-14B
Alibaba Group · None effort
53.4%IndependentEpoch AI28 Aug 2026
gpt-oss-20b
OpenAI · Low effort
53.2%IndependentEpoch AI27 Aug 2026
Qwen3-4B
Alibaba Group
52.3%IndependentEpoch AI28 Aug 2026
Llama 4 Scout
Meta
51.8%IndependentEpoch AI8 Apr 2025
Mistral Large 2.1
Mistral AI
51.3%IndependentEpoch AI25 Feb 2025
Llama 3.1 405B
Meta
50.9%IndependentEpoch AI27 Jan 2025
Command A
Cohere
50.5%Self-reportedEvalEval via Hugging Face Community Evals—
Qwen3-30B-A3B
Alibaba Group · None effort
50.4%IndependentEpoch AI30 Aug 2026
o1-preview
OpenAI
50.3%IndependentEpoch AI27 Jan 2025
GPT-4o
OpenAI
49.2%IndependentEpoch AI27 Jan 2025
Qwen2.5-72B-Instruct
Alibaba Group
49.1%IndependentEpoch AI27 Jan 2025
Mistral Large 2.0
Mistral AI
49.0%IndependentEpoch AI27 Jan 2025
GPT-4.1 nano
OpenAI
48.9%IndependentEpoch AI14 Apr 2025
GPT-4o
OpenAI
48.9%IndependentEpoch AI27 Jan 2025
GPT-5 nano
OpenAI · Minimal effort
48.5%IndependentEpoch AI13 Jul 2026
Qwen-Plus
Alibaba Group
48.1%IndependentEpoch AI7 Apr 2025
GPT-4o
OpenAI
47.9%IndependentEpoch AI5 Feb 2025
Gemma 3 27B
Google DeepMind
47.7%IndependentEpoch AI27 Aug 2026
Magistral Small 1.2
Mistral AI
47.6%IndependentEpoch AI30 Aug 2026
Mistral Small 3.1
Mistral AI
47.5%IndependentEpoch AI18 Mar 2025
Qwen3-8B
Alibaba Group · None effort
47.5%IndependentEpoch AI27 Aug 2026
Llama 3.3 70B
Meta
47.4%IndependentEpoch AI27 Jan 2025
Gemini 1.5 Flash
Google DeepMind
47.3%IndependentEpoch AI27 Jan 2025
Claude 3 Opus
Anthropic
47.2%IndependentEpoch AI27 Jan 2025
GPT-4 Turbo
OpenAI
46.6%IndependentEpoch AI27 Feb 2025
Llama 3.2 90B Vision
Meta
46.1%Self-reportedEvalEval via Hugging Face Community Evals—
Qwen2.5-32B-Instruct
Alibaba Group
46.1%IndependentEpoch AI30 Jan 2025
gpt-oss-20b
OpenAI · High effort
46.0%IndependentEpoch AI6 Aug 2026
Gemini 1.5 Pro
Google DeepMind
45.9%IndependentEpoch AI27 Jan 2025
Qwen3-4B-Instruct-2507
Alibaba Group
45.8%IndependentEpoch AI28 Aug 2026
Granite 4.1 30B
IBM
45.8%Self-reportedModel Card via Hugging Face Community Evals—
Mistral Small 3.0
Mistral AI
45.3%IndependentEpoch AI30 Jan 2025
GLM-4.7-Flash
Z.ai (Zhipu AI) · None effort
45.1%IndependentEpoch AI30 Aug 2026
DeepSeek-R1-Distill-Qwen-14B
DeepSeek
44.7%IndependentEpoch AI10 Mar 2025
Llama 3.1 70B
Meta
44.2%IndependentEpoch AI27 Jan 2025
Qwen3-4B
Alibaba Group · None effort
43.6%IndependentEpoch AI28 Aug 2026
WizardLM-2 8x22B
Microsoft
43.4%IndependentEpoch AI27 Jan 2025
GPT-4 Turbo
OpenAI
42.4%IndependentEpoch AI27 Jan 2025
GPT-4 Turbo
OpenAI
42.3%IndependentEpoch AI27 Jan 2025
Granite 4.1 8B
IBM
42.0%Self-reportedModel Card via Hugging Face Community Evals—
Qwen-Turbo
Alibaba Group
41.8%IndependentEpoch AI7 Apr 2025
Llama 3.2 90B Vision
Meta
41.0%IndependentEpoch AI27 Jan 2025
Qwen2-72B-Instruct
Alibaba Group
40.8%IndependentEpoch AI27 Jan 2025
Claude 3 Sonnet
Anthropic
40.6%IndependentEpoch AI27 Jan 2025
Llama 3 70B
Meta
40.6%IndependentEpoch AI27 Jan 2025
Gemini 1.5 Flash
Google DeepMind
40.4%IndependentEpoch AI27 Jan 2025
Gemma 3 12B
Google DeepMind
39.5%IndependentEpoch AI28 Aug 2026
Mistral Large 1.0
Mistral AI
38.8%IndependentEpoch AI27 Jan 2025
Claude 3.5 Haiku
Anthropic
38.1%IndependentEpoch AI12 Mar 2025
Qwen3-1.7B
Alibaba Group
38.0%IndependentEpoch AI28 Aug 2026
GPT-4o mini
OpenAI
37.7%IndependentEpoch AI27 Jan 2025
Hermes 2 Theta Llama 3 70B
Nous Research
37.5%IndependentEpoch AI27 Jan 2025
Gemma 2 27B
Google DeepMind
36.5%IndependentEpoch AI27 Jan 2025
Claude 3 Haiku
Anthropic
36.3%IndependentEpoch AI27 Jan 2025
GPT-4
OpenAI
35.7%IndependentEpoch AI23 Oct 2025
Qwen2.5-7B-Instruct
Alibaba Group
35.5%IndependentEpoch AI30 Aug 2026
Claude 2
Anthropic
34.7%IndependentEpoch AI27 Jan 2025
Command R+ (08-2024)
Cohere
34.3%Self-reportedEvalEval via Hugging Face Community Evals—
Mixtral 8x22B
Mistral AI
34.1%IndependentEpoch AI27 Jan 2025
Gemini 1.0 Pro
Google DeepMind
34.0%IndependentEpoch AI27 Jan 2025
DeepSeek-R1-Distill-Qwen-1.5B
DeepSeek
33.6%IndependentEpoch AI28 Aug 2026
Claude 2.1
Anthropic
33.0%IndependentEpoch AI27 Jan 2025
Gemini 1.5 Flash-8B
Google DeepMind
33.0%IndependentEpoch AI12 Mar 2025
DBRX Instruct
Databricks
32.9%IndependentEpoch AI27 Jan 2025
Yi-1.5-34B-Chat
01.AI
32.0%IndependentEpoch AI27 Jan 2025
Granite 4.1 3B
IBM
31.7%Self-reportedModel Card via Hugging Face Community Evals—
Qwen1.5-32B-Chat
Alibaba Group
30.7%IndependentEpoch AI27 Jan 2025
GPT-4
OpenAI
30.7%IndependentEpoch AI27 Jan 2025
Qwen3-1.7B
Alibaba Group · None effort
30.6%IndependentEpoch AI28 Aug 2026
Mixtral 8x7B
Mistral AI
30.6%IndependentEpoch AI27 Jan 2025
Mistral NeMo 12B
Mistral AI
29.9%IndependentEpoch AI27 Jan 2025
Mixtral 8x7B
Mistral AI
29.8%IndependentEpoch AI27 Jan 2025
Qwen1.5-72B-Chat
Alibaba Group
28.8%IndependentEpoch AI27 Jan 2025
Granite 4.0 Micro (3B)
IBM
28.3%IndependentEpoch AI30 Aug 2026
GPT-3.5 Turbo
OpenAI
28.0%IndependentEpoch AI27 Jan 2025
Phi-3-medium
Microsoft
27.6%IndependentEpoch AI31 Jan 2025
Gemma 2 9B
Google DeepMind
27.5%IndependentEpoch AI27 Jan 2025
GPT-3.5 Turbo
OpenAI
27.2%IndependentEpoch AI27 Jan 2025
Ministral 8B
Mistral AI
27.1%IndependentEpoch AI27 Jan 2025
Llama 3.1 8B
Meta
27.0%IndependentEpoch AI27 Aug 2026
Command R (08-2024)
Cohere
26.8%Self-reportedEvalEval via Hugging Face Community Evals—
Command R7B
Cohere
26.8%Self-reportedEvalEval via Hugging Face Community Evals—
Llama 2 70B
Meta
26.3%IndependentEpoch AI27 Jan 2025
Llama 3 8B
Meta
26.1%IndependentEpoch AI27 Jan 2025
Ministral 3B
Mistral AI
25.3%IndependentEpoch AI27 Jan 2025
DeepSeek LLM 67B Chat
DeepSeek
24.6%IndependentEpoch AI27 Jan 2025
Granite 4.0 1B (Nano)
IBM
24.0%IndependentEpoch AI28 Aug 2026
Llama 3.2 1B
Meta
23.9%IndependentEpoch AI30 Aug 2026
Gemma 3 4B
Google DeepMind
23.2%IndependentEpoch AI28 Aug 2026
Gemma 3 1B
Google DeepMind
19.9%IndependentEpoch AI28 Aug 2026
Llama 3.2 1B
Meta
18.7%Self-reportedEvalEval via Hugging Face Community Evals—
Mistral 7B v0.3
Mistral AI
15.2%IndependentEpoch AI27 Jan 2025
Yi-34B-Chat
01.AI
14.7%IndependentEpoch AI27 Jan 2025
Mistral 7B v0.3
Mistral AI
13.2%IndependentEpoch AI27 Jan 2025
Granite 4.0 350M (Nano)
IBM
11.2%IndependentEpoch AI28 Aug 2026
DeepSeek-R1-0528-Qwen3-8B
DeepSeek
9.3%IndependentEpoch AI28 Aug 2026
  • Epoch AI, 'Capabilities & benchmarking', epoch.ai/benchmarks, CC BY 4.0.
  • Reported on Hugging Face Community Evals; each result names its author and links its source.

Sources

GPQA Diamond: leaderboard and what it measures · supermarket.dev