MMLU-Pro
OpenCompass · 48 models
| Rank | Model | Score |
|---|---|---|
| 1 | Gemini-3-Pro-Preview | 89.3 |
| 2 | Qwen3.5-397B-A17B | 87.6 |
| 3 | GPT-5-2025-08-07 (high) | 86.2 |
| 3 | Kimi-K2.5 | 86.2 |
| 5 | DeepSeek-V3.2 | 85.8 |
| 5 | Gemini-2.5-Pro | 85.8 |
| 7 | JT-Think | 85.6 |
| 8 | DeepSeek-V3.2-Speciale | 85.5 |
| 9 | GLM-5 | 85.2 |
| 10 | Kimi-K2-Thinking | 84.3 |
| 11 | TeleChat-thinking | 84.1 |
| 12 | GLM-4.7 | 84 |
| 13 | DeepSeek-R1-0528 | 83.5 |
| 13 | Qwen3-235B-A22B-Thinking-2507 | 83.5 |
| 13 | Step-3.5-Flash | 83.5 |
| 16 | MiMo-V2-Flash | 83.1 |
| 16 | doubao-seed-1.6-thinking-250615 | 83.1 |
| 18 | Claude Sonnet 4 (Thinking) | 83 |
| 18 | GLM-4.6 | 83 |
| 18 | o3-2025-04-16 | 83 |
| 21 | Finix-P1-32B (Thinking) | 82.7 |
| 21 | GLM-4.5 | 82.7 |
| 23 | o4-mini-2025-04-16 (high) | 82.4 |
| 24 | Qwen3-Next-80B-A3B-Thinking | 82 |
| 25 | MiniMax-M2.5 | 81.7 |
| 26 | MiniMax-M2 | 81.6 |
| 27 | Qwen3-Next-80B-A3B-Instruct | 81.3 |
| 28 | GPT-4.1-20250414 | 81 |
| 28 | LongCat-Flash-Chat | 81 |
| 30 | Seed-OSS-36B-Instruct | 80.6 |
| 31 | gpt-oss-120b (high) | 79.7 |
| 32 | Llama4-Maverick-17B-128E-Instruct | 79.6 |
| 33 | Qwen3-30B-A3B-Thinking-2507 | 79.5 |
| 34 | Qwen3-235B-A22B-Instruct-2507 | 79.2 |
| 35 | ERNIE-4.5-300B-A47B-PT | 78.7 |
| 36 | Kimi-K2-Instruct | 78.4 |
| 37 | Qwen3-32B (Thinking) | 78 |
| 38 | GPT-5-nano-2025-08-07 (high) | 75 |
| 39 | Qwen3-30B-A3B-Instruct-2507 | 73.9 |
| 40 | Hunyuan-A13B-Instruct | 73.8 |
| 41 | Qwen3-4B-Thinking-2507 | 72.8 |
| 41 | gpt-oss-20b (high) | 72.8 |
| 43 | Qwen3-8B (Thinking) | 72.1 |
| 44 | MiniCPM-SALA | 72 |
| 45 | ERNIE-4.5-21B-A3B-Thinking | 70.8 |
| 46 | GPT-5-mini-2025-08-07 (high) | 70.2 |
| 47 | Gemma-3-27B-it | 67.8 |
| 48 | Qwen3-4B-Instruct-2507 | 63 |