AIME 2025
OpenCompass · 48 models
| Rank | Model | Score |
|---|---|---|
| 1 | DeepSeek-V3.2-Speciale | 96 |
| 2 | GLM-5 | 95.8 |
| 3 | Step-3.5-Flash | 95.7 |
| 4 | GLM-4.7 | 95.4 |
| 5 | GPT-5-2025-08-07 (high) | 94.1 |
| 5 | Kimi-K2-Thinking | 94.1 |
| 7 | Gemini-3-Pro-Preview | 93.4 |
| 7 | gpt-oss-120b (high) | 93.4 |
| 9 | DeepSeek-V3.2 | 93 |
| 10 | MiMo-V2-Flash | 92.9 |
| 11 | Qwen3.5-397B-A17B | 92.3 |
| 12 | Kimi-K2.5 | 91.9 |
| 13 | o4-mini-2025-04-16 (high) | 91 |
| 14 | GPT-5-mini-2025-08-07 (high) | 90.9 |
| 14 | Qwen3-235B-A22B-Thinking-2507 | 90.9 |
| 16 | GLM-4.6 | 90.3 |
| 17 | DeepSeek-R1-0528 | 89 |
| 17 | Qwen3-Next-80B-A3B-Thinking | 89 |
| 19 | Gemini-2.5-Pro | 88.7 |
| 20 | gpt-oss-20b (high) | 87.9 |
| 21 | Qwen3-30B-A3B-Thinking-2507 | 86.8 |
| 22 | MiniMax-M2.5 | 86.2 |
| 23 | Seed-OSS-36B-Instruct | 86 |
| 24 | GLM-4.5 | 85.8 |
| 25 | GPT-5-nano-2025-08-07 (high) | 85.4 |
| 26 | o3-2025-04-16 | 85.1 |
| 27 | doubao-seed-1.6-thinking-250615 | 82.8 |
| 28 | MiniCPM-SALA | 81 |
| 29 | Qwen3-4B-Thinking-2507 | 80 |
| 30 | MiniMax-M2 | 79.1 |
| 31 | TeleChat-thinking | 79 |
| 32 | ERNIE-4.5-21B-A3B-Thinking | 76.2 |
| 33 | Qwen3-32B (Thinking) | 70.3 |
| 34 | Finix-P1-32B (Thinking) | 70.2 |
| 35 | Qwen3-235B-A22B-Instruct-2507 | 69.5 |
| 36 | Qwen3-Next-80B-A3B-Instruct | 69.2 |
| 37 | Claude Sonnet 4 (Thinking) | 68.7 |
| 38 | Qwen3-8B (Thinking) | 66.2 |
| 39 | JT-Think | 65.9 |
| 40 | Hunyuan-A13B-Instruct | 65.7 |
| 41 | Qwen3-30B-A3B-Instruct-2507 | 63.8 |
| 42 | LongCat-Flash-Chat | 61 |
| 43 | Kimi-K2-Instruct | 49.1 |
| 44 | Qwen3-4B-Instruct-2507 | 46.9 |
| 45 | GPT-4.1-20250414 | 35.3 |
| 46 | ERNIE-4.5-300B-A47B-PT | 32.6 |
| 47 | Gemma-3-27B-it | 22.4 |
| 48 | Llama4-Maverick-17B-128E-Instruct | 15.2 |