GPQA Diamond
OpenCompass · 48 models
| Rank | Model | Score |
|---|---|---|
| 1 | Gemini-3-Pro-Preview | 91.5 |
| 2 | Qwen3.5-397B-A17B | 88.4 |
| 3 | Kimi-K2.5 | 88.1 |
| 4 | GLM-4.7 | 86.9 |
| 5 | DeepSeek-V3.2-Speciale | 86.7 |
| 6 | GPT-5-2025-08-07 (high) | 86.4 |
| 7 | GLM-5 | 85.3 |
| 8 | Gemini-2.5-Pro | 84.7 |
| 9 | DeepSeek-V3.2 | 84.6 |
| 9 | MiniMax-M2.5 | 84.6 |
| 11 | Step-3.5-Flash | 83.7 |
| 12 | ERNIE-4.5-300B-A47B-PT | 83 |
| 13 | GPT-5-mini-2025-08-07 (high) | 82.7 |
| 13 | Kimi-K2-Thinking | 82.7 |
| 15 | o3-2025-04-16 | 82.3 |
| 16 | MiMo-V2-Flash | 82.1 |
| 17 | TeleChat-thinking | 81.4 |
| 18 | DeepSeek-R1-0528 | 80.6 |
| 19 | GLM-4.6 | 80.4 |
| 20 | Qwen3-235B-A22B-Thinking-2507 | 79.8 |
| 21 | GLM-4.5 | 79.5 |
| 22 | gpt-oss-120b (high) | 78.9 |
| 23 | MiniMax-M2 | 78.7 |
| 24 | JT-Think | 77.7 |
| 25 | o4-mini-2025-04-16 (high) | 77.4 |
| 26 | Qwen3-Next-80B-A3B-Thinking | 77 |
| 27 | doubao-seed-1.6-thinking-250615 | 75.9 |
| 28 | Qwen3-235B-A22B-Instruct-2507 | 75.5 |
| 29 | Claude Sonnet 4 (Thinking) | 74.6 |
| 30 | Qwen3-Next-80B-A3B-Instruct | 74.1 |
| 31 | Kimi-K2-Instruct | 72.7 |
| 32 | Seed-OSS-36B-Instruct | 71.5 |
| 33 | LongCat-Flash-Chat | 70.1 |
| 33 | Qwen3-30B-A3B-Thinking-2507 | 70.1 |
| 35 | Llama4-Maverick-17B-128E-Instruct | 69.1 |
| 36 | gpt-oss-20b (high) | 68.9 |
| 37 | Finix-P1-32B (Thinking) | 68.7 |
| 38 | GPT-4.1-20250414 | 68.1 |
| 39 | GPT-5-nano-2025-08-07 (high) | 67.8 |
| 40 | Qwen3-32B (Thinking) | 67.3 |
| 41 | Qwen3-4B-Thinking-2507 | 64.7 |
| 42 | Hunyuan-A13B-Instruct | 64.1 |
| 43 | Qwen3-30B-A3B-Instruct-2507 | 63.6 |
| 44 | ERNIE-4.5-21B-A3B-Thinking | 60.2 |
| 45 | Qwen3-8B (Thinking) | 59.7 |
| 46 | MiniCPM-SALA | 58.2 |
| 47 | Qwen3-4B-Instruct-2507 | 52.3 |
| 48 | Gemma-3-27B-it | 46.3 |