| # | 模型 | 分数 |
|---|---|---|
| 1 | DeepSeek V4.1-Flash 1017 | 60.5% |
| 2 | DeepSeek V4.1-Flash 0910 | 54.8% |
| 3 | Gemini 4 Argon | 51.3% |
| 4 | Claude Opus 5 | 50.3% |
| 5 | GLM 5.3 | 48.8% |
| 6 | Kimi K3 | 46.7% |
| 7 | GPT 5.6-Sol | 45.8% |
| 8 | DeepSeek V4-Pro 0813 | 43.2% |
| 9 | DeepSeek V4-Flash 0731 | 37.7% |
| 10 | Gemini 3.1 Pro | 35.4% |