GPQA Diamond
GPQA Diamond is run by Epoch AI (AI Benchmarking Hub). It has ranked 326 models, of which the catalogue holds 185, scoring from 9.28 to 96.402 on Accuracy.
measured by Epoch AI (AI Benchmarking Hub) · the board itself
| Place | Model | Metric | Score |
|---|---|---|---|
| 1stof 326 | Gemini 4 Argon | mean_score (high) | 96.402 |
| 2ndof 326 | GPT 6 Astra | mean_score (max) | 95.77 |
| 3rdof 326 | Claude Sonnet 5.5 | mean_score (max) | 95.581 |
| 4thof 326 | GPT 6.1 Sol | mean_score (max) | 95.391 |
| 5thof 326 | Gemini 3.8 Flash | mean_score (high) | 95.391 |
| 6thof 326 | Gemini 3.7 Flash | mean_score (high) | 94.823 |
| 7thof 326 | GPT-5.4 Pro | mean_score (xhigh) | 94.6 |
| 8thof 326 | Gemini 3.1 Pro Preview | mean_score (high) | 94.444 |
| 9thof 326 | GPT-6.1 Terra | mean_score (max) | 94.318 |
| 10thof 326 | GPT 6 Sol | mean_score (max) | 94.318 |
| 11thof 326 | Gemini 3.6 Flash | mean_score (high) | 94.129 |
| 13thof 326 | Grok 4.6 | mean_score (high) | 94.003 |
| 14thof 326 | GPT-5.5 | mean_score (xhigh) | 94.003 |
| 15thof 326 | GPT-5.5 Pro | mean_score (xhigh) | 93.921 |
| 16thof 326 | Claude Opus 5 | mean_score (max) | 93.876 |
| 17thof 326 | GPT-5.6 Sol | mean_score (max) | 93.497 |
| 18thof 326 | Grok 4.5 | mean_score (high) | 93.434 |
| 19thof 326 | GPT-5.6 Terra | mean_score (max) | 93.308 |
| 20thof 326 | GPT-5.4 | mean_score (xhigh) | 93.3 |
| 21stof 326 | DeepSeek V4.1 Flash 1017 | mean_score (max) | 93.182 |
| 23rdof 326 | Kimi K3 | mean_score (max) | 93.119 |
| 25thof 326 | Gemini 3.5 Flash | mean_score (high) | 92.803 |
| 26thof 326 | Grok 4.7 | mean_score (xhigh) | 92.74 |
| 27thof 326 | Qwen 3.8 Max | mean_score (xhigh) | 92.677 |
| 28thof 326 | gemini-3-pro-preview | mean_score | 92.614 |
| 31stof 326 | GLM 5.2 | mean_score (max) | 91.856 |
| 32ndof 326 | DeepSeek V4 Pro 0813 | mean_score (max) | 91.667 |
| 33rdof 326 | GPT-5.6 Luna | mean_score (max) | 91.604 |
| 34thof 326 | GPT-5.2 | mean_score (xhigh) | 91.4 |
| 35thof 326 | DeepSeek V4 Flash (0731) | mean_score (max) | 91.035 |
| 36thof 326 | Claude Opus 4.8 | mean_score (max) | 91.035 |
| 37thof 326 | GLM-5.3 | mean_score (max) | 90.909 |
| 38thof 326 | MiniMax M3 | mean_score | 90.909 |
| 39thof 326 | Qwen3.7 Max | mean_score (max) | 90.909 |
| 40thof 326 | DeepSeek V4 Pro | mean_score (high) | 90.909 |
| 41stof 326 | Kimi K2.6 | mean_score | 90.783 |
| 43rdof 326 | Claude Opus 5.5 | mean_score (max) | 90.593 |
| 44thof 326 | Claude Sonnet 5 | mean_score (xhigh) | 90.53 |
| 45thof 326 | Claude Opus 4.6 | mean_score | 90.53 |
| 46thof 326 | GPT 6 Luna | mean_score (max) | 90.467 |
| 47thof 326 | GLM 5.3 Flash | mean_score (max) | 90.152 |
| 48thof 326 | Claude Opus 4.7 | mean_score (xhigh) | 90.152 |
| 49thof 326 | GLM 5.1 | mean_score | 89.899 |
| 52ndof 326 | DeepSeek V4.1 Flash | mean_score (max) | 89.836 |
| 53rdof 326 | Muse Spark 1.2 | mean_score | 89.8 |
| 54thof 326 | Claude Haiku 5.5 | mean_score (max) | 89.646 |
| 56thof 326 | Gemini 3 Flash | mean_score (high) | 89.394 |
| 57thof 326 | Grok 4.20 | mean_score | 89.331 |
| 60thof 326 | Grok 4.3 | mean_score (high) | 88.826 |
| 62ndof 326 | Muse-Spark-Open-816B-A42B | mean_score (xhigh) | 88.573 |
| 63rdof 326 | Inkling Small | mean_score (xhigh) | 88.51 |
| 64thof 326 | Qwen3.6 Plus | mean_score | 88.384 |
| 67thof 326 | Inkling | mean_score (xhigh) | 88.258 |
| 71stof 326 | Kimi K2.7 Code | mean_score | 87.879 |
| 72ndof 326 | Qwen 3.7 Plus | mean_score | 87.879 |
| 75thof 326 | GLM-5 | mean_score | 87.817 |
| 76thof 326 | GPT-5.1 | mean_score (high) | 87.626 |
| 77thof 326 | Kimi K2.5 | mean_score | 87.6 |
| 78thof 326 | Qwen3.6 Max Preview | mean_score (max) | 87.374 |
| 80thof 326 | Claude Sonnet 4.6 | mean_score | 87.374 |
| 82ndof 326 | GPT-5.4 mini | mean_score (xhigh) | 86.869 |
| 83rdof 326 | Qwen3.5 397B A17B | mean_score (none) | 86.364 |
| 87thof 326 | GPT-5 | mean_score (high) | 86.174 |
| 88thof 326 | Claude Opus 4.5 | mean_score | 86.048 |
| 89thof 326 | Qwen3.6 27B | mean_score | 85.859 |
| 92ndof 326 | Claude Fable 5 | mean_score (max) | 85.859 |
| 94thof 326 | Nemotron 3 Ultra | mean_score | 85.354 |
| 97thof 326 | Gemini 2.5 Pro | mean_score | 85.29 |
| 101stof 326 | Qwen 3.5 Plus | mean_score | 84.848 |
| 102ndof 326 | Qwen3.6 35B A3B | mean_score (none) | 84.848 |
| 104thof 326 | Gemini 2.5 Pro Preview 06-05 | mean_score | 84.848 |
| 109thof 326 | Qwen3.5-35B-A3B | mean_score | 83.46 |
| 110thof 326 | deepseek-reasoner | mean_score | 83.424 |
| 111thof 326 | Qwen3.6-Flash | mean_score | 83.333 |
| 112thof 326 | Gemini 3.5 Flash-Lite | mean_score (high) | 83.333 |
| 116thof 326 | GLM-4.7 | mean_score | 83.333 |
| 117thof 326 | Gemini 3 Flash Preview | mean_score | 83.207 |
| 120thof 326 | GPT-5.5 Instant | mean_score | 82.513 |
| 122ndof 326 | Qwen3.7 Flash | mean_score | 82.323 |
| 123rdof 326 | Qwen3.5-Flash | mean_score | 82.323 |
| 124thof 326 | Claude Sonnet 4.5 | mean_score | 82.323 |
| 126thof 326 | Gemini 3.1 Flash-Lite | mean_score (high) | 81.818 |
| 127thof 326 | o3 | mean_score (high) | 81.818 |
| 135thof 326 | Qwen3 235B A22B Thinking 2507 | mean_score | 80.051 |
| 137thof 326 | o4-mini | mean_score (high) | 79.609 |
| 138thof 326 | Qwen3.5 9B | mean_score | 78.977 |
| 143rdof 326 | Claude 3.7 Sonnet | mean_score | 78.504 |
| 144thof 326 | GPT-5.4 nano | mean_score (high) | 78.47 |
| 145thof 326 | Claude Sonnet 4 | mean_score | 78.283 |
| 150thof 326 | Claude Opus 4.1 | mean_score | 77.273 |
| 151stof 326 | o3-mini | mean_score (high) | 77.02 |
| 155thof 326 | o1 | mean_score (high) | 76.768 |
| 156thof 326 | DeepSeek-R1 (0528) | mean_score | 76.326 |
| 157thof 326 | Claude Opus 4 | mean_score | 76.263 |
| 158thof 326 | Grok 3 Mini | mean_score (low) | 76.263 |
| 159thof 326 | Gemma 4 31B | mean_score (minimal) | 75.758 |
| 161stof 326 | GPT OSS 120B | mean_score (high) | 75.758 |
| 165thof 326 | GPT-5 mini | mean_score (high) | 75 |
| 183rdof 326 | Seed-OSS-36B-Instruct | mean_score | 71.528 |
| 185thof 326 | deepseek-chat | mean_score | 71.212 |
| 186thof 326 | Claude Haiku 4.5 | mean_score | 71.212 |
| 187thof 326 | Qwen3 235B A22B | mean_score | 70.707 |
| 188thof 326 | Qwen3 30B A3B Thinking 2507 | mean_score | 70.076 |
| 189thof 326 | GPT-5 nano | mean_score (high) | 69.444 |
| 190thof 326 | DeepSeek-R1 | mean_score | 69.223 |
| 194thof 326 | DeepSeek-V3 0324 | mean_score | 67.614 |
| 195thof 326 | Grok 3 | mean_score | 67.582 |
| 197thof 326 | Llama 4 Maverick 17B | mean_score | 66.982 |
| 198thof 326 | GPT-4.1 fine-tuned | mean_score | 66.919 |
| 200thof 326 | Gemini 2.5 Pro Preview 05-06 | mean_score | 66.667 |
| 203rdof 326 | GPT-4.1 mini | mean_score | 65.846 |
| 204thof 326 | Qwen3 32B | mean_score | 65.72 |
| 206thof 326 | QwQ-Plus | mean_score | 65.404 |
| 207thof 326 | QwQ-32B | mean_score | 65.341 |
| 208thof 326 | DeepSeek R1 Distill QWEN 32B | mean_score | 64.141 |
| 210thof 326 | gemini-2.0-flash-001 | mean_score | 64.141 |
| 211thof 326 | Qwen3 14B | mean_score | 63.763 |
| 213thof 326 | o1-mini | mean_score (high) | 62.374 |
| 214thof 326 | Qwen3 30B A3B | mean_score | 61.679 |
| 215thof 326 | GPT OSS 20B | mean_score (medium) | 60.795 |
| 216thof 326 | GLM-4.7 Flash | mean_score | 60.543 |
| 218thof 326 | Mistral Medium 2505 | mean_score | 59.533 |
| 223rdof 326 | Qwen3 8B | mean_score | 56.755 |
| 224thof 326 | DeepSeek-V3 | mean_score | 56.534 |
| 226thof 326 | Magistral Small 2506 | mean_score | 56.061 |
| 227thof 326 | Phi-4 | mean_score | 56.061 |
| 228thof 326 | R1 Distill Llama 70B | mean_score | 55.745 |
| 229thof 326 | Qwen3 30B A3B Instruct 2507 | mean_score | 55.619 |
| 231stof 326 | Claude 3.5 Sonnet | mean_score | 55.303 |
| 237thof 326 | Qwen3 4B | mean_score | 52.336 |
| 240thof 326 | Meta-Llama-3.1-405B-Instruct | mean_score | 50.915 |
| 243rdof 326 | GPT-4o (2024-08-06) | mean_score | 49.211 |
| 244thof 326 | Qwen2.5 72B Instruct | mean_score | 49.148 |
| 246thof 326 | Mistral Large 2407 | mean_score | 49.021 |
| 247thof 326 | GPT-4.1 nano | mean_score | 48.927 |
| 248thof 326 | GPT-4o (2024-05-13) | mean_score | 48.895 |
| 250thof 326 | Qwen-Plus | mean_score | 48.106 |
| 251stof 326 | GPT-4o (2024-11-20) | mean_score | 47.885 |
| 252ndof 326 | Gemma 3 27B | mean_score | 47.727 |
| 255thof 326 | mistral-small-2503 | mean_score | 47.475 |
| 256thof 326 | Llama 3.3 70B Instruct | mean_score | 47.443 |
| 259thof 326 | Claude 3 Opus | mean_score | 47.159 |
| 260thof 326 | GPT-4 Turbo | mean_score | 46.591 |
| 262ndof 326 | AgentRL | mean_score | 46.086 |
| 265thof 326 | Qwen3-4B-Instruct-2507 | mean_score | 45.833 |
| 268thof 326 | DeepSeek R1 Distill QWEN 14B | mean_score | 44.697 |
| 269thof 326 | Meta-Llama-3.1-70B-Instruct | mean_score | 44.192 |
| 271stof 326 | WizardLM-2 8x22B | mean_score | 43.434 |
| 274thof 326 | Mistral Small 3.1 (25.03) | mean_score | 41.919 |
| 275thof 326 | qwen-turbo | mean_score | 41.793 |
| 276thof 326 | Llama-3.2-90B-Vision-Instruct | mean_score | 41.035 |
| 277thof 326 | Qwen2-72B-Instruct | mean_score | 40.783 |
| 278thof 326 | Claude 3 Sonnet | mean_score | 40.593 |
| 279thof 326 | Meta-Llama-3-70B-Instruct | mean_score | 40.562 |
| 281stof 326 | Gemma 3 12B | mean_score | 39.457 |
| 282ndof 326 | Mistral Large (24.02) | mean_score | 38.763 |
| 283rdof 326 | Claude Haiku 3.5 | mean_score | 38.131 |
| 284thof 326 | Qwen3-1.7B | mean_score | 38.005 |
| 285thof 326 | GPT-4o-mini (2024-07-18) | mean_score | 37.721 |
| 287thof 326 | Gemma 2 27B | mean_score | 36.49 |
| 288thof 326 | Claude 3 Haiku | mean_score | 36.301 |
| 290thof 326 | Qwen2.5 7B Instruct | mean_score | 35.48 |
| 291stof 326 | Claude 2.0 | mean_score | 34.659 |
| 295thof 326 | DeepSeek-R1-Distill-Qwen-1.5B | mean_score | 33.586 |
| 297thof 326 | Claude 2.1 | mean_score | 32.955 |
| 299thof 326 | Yi-1.5-34B-Chat | mean_score | 31.976 |
| 301stof 326 | GPT-4 (0613) | mean_score | 30.65 |
| 303rdof 326 | Mixtral-8x7B-Instruct-v0.1 | mean_score | 30.587 |
| 306thof 326 | Qwen1.5-72B-Chat | mean_score | 28.819 |
| 307thof 326 | Granite 4.0 Micro | mean_score | 28.283 |
| 308thof 326 | GPT-3.5 Turbo (1106) | mean_score | 28.03 |
| 309thof 326 | Phi-3-medium-128k-instruct | mean_score | 27.588 |
| 310thof 326 | gemma-2-9b-it | mean_score | 27.462 |
| 311thof 326 | GPT-3.5 Turbo | mean_score | 27.178 |
| 313thof 326 | Llama 3.1 8B Instruct | mean_score | 26.957 |
| 314thof 326 | Llama-2-70b-chat-hf | mean_score | 26.326 |
| 315thof 326 | Llama 3 8B Instruct | mean_score | 26.073 |
| 317thof 326 | deepseek-llm-67b-chat | mean_score | 24.621 |
| 319thof 326 | Llama 3.2 1B Instruct | mean_score | 23.927 |
| 320thof 326 | Gemma 3 4B | mean_score | 23.232 |
| 321stof 326 | gemma-3-1b-it | mean_score | 19.949 |
| 322ndof 326 | Mistral-7B-Instruct-v0.3 | mean_score | 15.183 |
| 323rdof 326 | Yi-34B-Chat | mean_score | 14.741 |
| 324thof 326 | open-mistral-7b | mean_score | 13.226 |
| 326thof 326 | DeepSeek R1 0528 Qwen3 8B | mean_score | 9.28 |
Read from the board on 2026-10-31