Mathematical Reasoning on AIME 25 (Average Rank Metrics)
1.83Average RankDIVER
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| DIVER2025.09 | 1.83 | 60 | 13.3 | 16.7 | 3.3 | 6.7 | |
| Pass@k Training2025.09 | 2.7 | 26.7 | 20 | 26.7 | 10 | 16.7 | |
| GPRO w/ Clip-Higher2025.09 | 2.83 | 16.7 | 26.7 | 23.3 | 23.3 | 10 | |
| Entropy-RL2025.09 | 3 | 20 | 13.3 | 23.3 | 33.3 | 10 | |
| Qwen2.5-Math-7B2025.09 | 3.73 | 0 | 30 | 10 | 16.7 | 43.3 |