GPQA Diamond (Vals run)
Graduate-level biology, chemistry and physics multiple-choice questions. Near saturation; shares weight with other GPQA Diamond runs. Run by Vals AI.
| Category | Reasoning |
|---|---|
| Metric | Accuracy (%) |
| Source | Vals AI (evaluator) |
| License | None stated |
| Benchmark page | www.vals.ai |
| Status | Scored |
| Weight | 0.8 within Reasoning |
| Fit | Discrimination 0.86, noise 0.26 |
| Lineage | Shares weight with gpqa_diamond |
Results
| # | Model | Provider | Result | Date |
|---|---|---|---|---|
| 1 | Gemini 3.1 Pro Preview High | 95.5% | 2026-09-01 | |
| 2 | GPT-5.6 Sol Max | OpenAI | 95.2% | 2026-09-01 |
| 3 | Grok 4.6 High | xAI | 94.7% | 2026-09-01 |
| 4 | Gemini 3.8 Flash High | 94.4% | 2026-09-01 | |
| 5 | Gemini 3.7 Flash High | 93.9% | 2026-09-01 | |
| 6 | Qwen3.8 Max | Alibaba | 93.7% | 2026-09-01 |
| 7 | Claude Opus 5 | Anthropic | 93.4% | 2026-09-01 |
| 8 | Gemini 3.6 Flash High | 93.4% | 2026-09-01 | |
| 9 | Claude Fable 5.1 Max | Anthropic | 93.4% | 2026-09-01 |
| 10 | Claude Fable 5 Max | Anthropic | 93.2% | 2026-09-01 |
| 11 | GPT-5.5 Extra High | OpenAI | 93.2% | 2026-09-01 |
| 12 | Kimi K3 | Moonshot AI | 92.9% | 2026-09-01 |
| 13 | Grok 4.5 High | xAI | 92.9% | 2026-09-01 |
| 14 | MiniMax-M3 | MiniMax | 92.7% | 2026-09-01 |
| 15 | Gemini 3.5 Flash High | 92.7% | 2026-09-01 | |
| 16 | DeepSeek V4 Pro 0813 Max | DeepSeek | 92.4% | 2026-09-01 |
| 17 | Claude Opus 4.8 Max | Anthropic | 92.4% | 2026-09-01 |
| 18 | GPT-5.6 Luna Max | OpenAI | 91.7% | 2026-09-01 |
| 19 | GPT-5.4 Extra High | OpenAI | 91.7% | 2026-09-01 |
| 20 | GPT-5.2 Extra High | OpenAI | 91.7% | 2026-09-01 |
| 21 | Gemini 3 Pro Preview High | 91.7% | 2026-09-01 | |
| 22 | Grok 4.3 | xAI | 91.4% | 2026-09-01 |
| 23 | Muse Spark 1.1 Extra High | Meta | 91.2% | 2026-09-01 |
| 24 | GPT-5.6 Terra Extra High | OpenAI | 90.9% | 2026-09-01 |
| 25 | Qwen3.7 Max | Alibaba | 90.2% | 2026-09-01 |
| 26 | Claude Opus 4.7 Max | Anthropic | 90.2% | 2026-09-01 |
| 27 | DeepSeek V4 Flash 0731 High | DeepSeek | 89.9% | 2026-09-01 |
| 28 | Muse Spark | Meta | 89.6% | 2026-09-01 |
| 29 | Claude Opus 4.6 Thinking | Anthropic | 89.6% | 2026-09-01 |
| 30 | DeepSeek V4 Pro Max | DeepSeek | 89.4% | 2026-09-01 |
| 31 | Kimi K2.6 | Moonshot AI | 89.1% | 2026-09-01 |
| 32 | Claude Sonnet 5 Max | Anthropic | 88.9% | 2026-09-01 |
| 33 | Qwen3.8 27B Extra High | Alibaba | 88.9% | 2026-09-01 |
| 34 | Grok 4.20 (Reasoning) | xAI | 88.6% | 2026-09-01 |
| 35 | Grok 4 | xAI | 88.1% | 2026-09-01 |
| 36 | GLM-5.3 Max | Zhipu AI | 88.1% | 2026-09-01 |
| 37 | Gemini 3 Flash Preview High | 87.9% | 2026-09-01 | |
| 38 | Qwen3.6 Plus | Alibaba | 87.4% | 2026-09-01 |
| 39 | Qwen3.5 Plus Thinking | Alibaba | 87.4% | 2026-09-01 |
| 40 | Inkling | Thinking Machines | 87.1% | 2026-09-01 |
| 41 | MiniMax-M2.7 | MiniMax | 86.6% | 2026-09-01 |
| 42 | GPT-5.1 High | OpenAI | 86.6% | 2026-09-01 |
| 43 | GLM-5.3-Flash Max | Zhipu AI | 86.4% | 2026-09-01 |
| 44 | Nemotron 3 Ultra 550B A55B | NVIDIA | 86.1% | 2026-09-01 |
| 45 | Claude Opus 4.5 Thinking | Anthropic | 85.9% | 2026-09-01 |
| 46 | GLM-5.2 | Zhipu AI | 85.6% | 2026-09-01 |
| 47 | Claude Sonnet 4.6 Max | Anthropic | 85.6% | 2026-09-01 |
| 48 | GPT-5 High | OpenAI | 85.6% | 2026-09-01 |
| 49 | X-Ai/Grok-4-Fast-Reasoning | xAI | 85.4% | 2026-09-01 |
| 50 | Qwen3 Max | Alibaba | 84.8% | 2026-09-01 |
| 51 | GLM-5.1 | Zhipu AI | 84.5% | 2026-09-01 |
| 52 | Grok 4.1 Fast (Reasoning) | xAI | 84.3% | 2026-09-01 |
| 53 | o3 High | OpenAI | 84.1% | 2026-09-01 |
| 54 | Kimi K2.5 Thinking | Moonshot AI | 84.1% | 2026-09-01 |
| 55 | Gemini 3.5 Flash Lite High | 83.8% | 2026-09-01 | |
| 56 | Inkling Small | Thinking Machines | 83.6% | 2026-09-01 |
| 57 | GLM-5 Thinking | Zhipu AI | 83.3% | 2026-09-01 |
| 58 | GPT-5.4 mini Extra High | OpenAI | 83.1% | 2026-09-01 |
| 59 | Qwen3.5 Flash | Alibaba | 82.8% | 2026-09-01 |
| 60 | MiMo-V2.5-Pro | Xiaomi | 82.6% | 2026-09-01 |
| 61 | MiniMax-M2.5 | MiniMax | 82.1% | 2026-09-01 |
| 62 | Claude Sonnet 4.5 Thinking | Anthropic | 81.6% | 2026-09-01 |
| 63 | MiMo-V2.5 | Xiaomi | 81.6% | 2026-09-01 |
| 64 | Gemini 2.5 Flash (Sep 2025) | 81.6% | 2026-09-01 | |
| 65 | Gemini 3.1 Flash Lite Preview High | 81.1% | 2026-09-01 | |
| 66 | Gemini 2.5 Pro Preview 06-05 | 80.8% | 2026-09-01 | |
| 67 | GPT-5 Mini High | OpenAI | 80.3% | 2026-09-01 |
| 68 | DeepSeek V3.2 High | DeepSeek | 80.3% | 2026-09-01 |
| 69 | GLM-4.7 | Zhipu AI | 80.1% | 2026-09-01 |
| 70 | Claude Opus 4.5 High | Anthropic | 79.5% | 2026-09-01 |
| 71 | Kimi K2 Thinking | Moonshot AI | 78.5% | 2026-09-01 |
| 72 | MiniMax-M2.1 | MiniMax | 78.5% | 2026-09-01 |
| 73 | GPT OSS 120B | OpenAI | 78.5% | 2026-09-01 |
| 74 | Qwen3 Max Preview | Alibaba | 77.8% | 2026-09-01 |
| 75 | GPT-5.4 nano High | OpenAI | 77.5% | 2026-09-01 |
| 76 | Gemini 2.5 Flash (Sep 2025) Thinking | 76.5% | 2026-09-01 | |
| 77 | Claude Opus 4.1 Thinking | Anthropic | 76.3% | 2026-09-01 |
| 78 | DeepSeek V3.2 Non-reasoning | DeepSeek | 76.3% | 2026-09-01 |
| 79 | o3-mini High | OpenAI | 75.5% | 2026-09-01 |
| 80 | Claude Sonnet 3.7 Thinking | Anthropic | 75.3% | 2026-09-01 |
| 81 | Claude Sonnet 4 Thinking | Anthropic | 75.0% | 2026-09-01 |
| 82 | GLM-4.6 | Zhipu AI | 74.5% | 2026-09-01 |
| 83 | o4-mini High | OpenAI | 74.5% | 2026-09-01 |
| 84 | Grok 3 | xAI | 74.2% | 2026-09-01 |
| 85 | o1 High | OpenAI | 73.2% | 2026-09-01 |
| 86 | GLM-4.5 | Zhipu AI | 72.2% | 2026-09-01 |
| 87 | Claude Haiku 4.5 Thinking | Anthropic | 72.2% | 2026-09-01 |
| 88 | Claude Opus 4 | Anthropic | 71.7% | 2026-09-01 |
| 89 | Kimi K2 Instruct | Moonshot AI | 71.5% | 2026-09-01 |
| 90 | Qwen3 235B-A22B | Alibaba | 70.2% | 2026-09-01 |
| 91 | Gemini 2.5 Flash-Lite Preview 09-2025 Thinking | 70.2% | 2026-09-01 | |
| 92 | Claude Opus 4.1 | Anthropic | 70.0% | 2026-09-01 |
| 93 | Claude Sonnet 4 | Anthropic | 69.4% | 2026-09-01 |
| 94 | GPT OSS 20B | OpenAI | 68.9% | 2026-09-01 |
| 95 | Mistral Large 3 | Mistral | 68.4% | 2026-09-01 |
| 96 | GPT-4.1 mini High | OpenAI | 67.9% | 2026-09-01 |
| 97 | Claude Sonnet 3.7 | Anthropic | 67.4% | 2026-09-01 |
| 98 | GPT-4.1 High | OpenAI | 65.4% | 2026-09-01 |
| 99 | Grok 4.1 Fast | xAI | 65.2% | 2026-09-01 |
| 100 | Gemini 2.0 Flash | 65.2% | 2026-09-01 | |
| 101 | Gemini 2.5 Flash-Lite Preview 09-2025 | 64.1% | 2026-09-01 | |
| 102 | GPT-5 Nano High | OpenAI | 63.4% | 2026-09-01 |
| 103 | Magistral Medium 1.2 | Mistral | 62.4% | 2026-09-01 |
| 104 | X-Ai/Grok-4-Fast-Non-Reasoning | xAI | 62.1% | 2026-09-01 |
| 105 | DeepSeek V3 0324 | DeepSeek | 61.6% | 2026-09-01 |
| 106 | Claude Sonnet 3.5 v2 | Anthropic | 59.3% | 2026-09-01 |
| 107 | MiMo-V2-Flash | Xiaomi | 59.3% | 2026-09-01 |
| 108 | Magistral Small 1.2 | Mistral | 58.3% | 2026-09-01 |
| 109 | Gemini 1.5 Pro (Sept 2024) | 58.3% | 2026-09-01 | |
| 110 | Gemini 2.5 Flash (Apr 2025) | 57.6% | 2026-09-01 | |
| 111 | Laguna XS.2 | Poolside | 55.1% | 2026-09-01 |
| 112 | DeepSeek-V3 | DeepSeek | 54.5% | 2026-09-01 |
| 113 | GPT-4o (2024-11-20) High | OpenAI | 53.8% | 2026-09-01 |
| 114 | Grok-2 (Dec 2024) | xAI | 50.8% | 2026-09-01 |
| 115 | mistral-small-2402 | Unknown | 50.8% | 2026-09-01 |
| 116 | GPT-4.1 nano High | OpenAI | 50.8% | 2026-09-01 |
| 117 | GPT-4o (2024-05-13) High | OpenAI | 50.3% | 2026-09-01 |
| 118 | Command A | Cohere | 48.5% | 2026-09-01 |
| 119 | Mistral Large 2.1 | Mistral | 47.7% | 2026-09-01 |
| 120 | Gemini 1.5 Flash (Sep 2024) | 46.0% | 2026-09-01 | |
| 121 | Mistral Small 3.1 | Mistral | 44.2% | 2026-09-01 |
| 122 | GPT-4o mini High | OpenAI | 44.2% | 2026-09-01 |
| 123 | Claude Haiku 3.5 | Anthropic | 37.9% | 2026-09-01 |
| 124 | Mistral Medium 3.5 High | Mistral | 34.8% | 2026-09-01 |
| 125 | GPT-3.5-turbo High | OpenAI | 30.6% | 2026-09-01 |
| 126 | Laguna M.1 | Poolside | 27.0% | 2026-09-01 |