Mathematical Reasoning on AIME 2026 (test)
16.7Majority Success Rate @32AER
Evaluation results
| Method | Links | ||
|---|---|---|---|
| AERK=322026.06 | 16.7 | 9.1 | |
| PAECK=322026.06 | 16.7 | 9.3 | |
| CISPOK=322026.06 | 13.3 | 8.9 | |
| Qwen2.5-Math-1.5BK=322026.06 | 10 | 6.3 | |
| GRPOK=322026.06 | 10 | 6.3 | |
| KL-covK=322026.06 | 10 | 7.1 | |
| Global Entropy RegK=322026.06 | 9.8 | 9 | |
| Clip-covK=322026.06 | 9.5 | 6.5 |