Mathematical Reasoning on AIME 24 (Pass@32)
62.4Pass@32OPEFO
Evaluation Results
| Method | Links | |
|---|---|---|
| OPEFOBackbone=Qwen2.5-Math-7B2026.05 | 62.4 | |
| GRPO (Strict)Backbone=Qwen2.5-Math-7B2026.05 | 61.3 | |
| KL-CovBackbone=Qwen2.5-Math-7B2026.05 | 60.1 | |
| GRPOBackbone=Qwen2.5-Math-7B2026.05 | 59.8 | |
| Clip-CovBackbone=Qwen2.5-Math-7B2026.05 | 59 | |
| Clip-higherBackbone=Qwen2.5-Math-7B2026.05 | 57.7 | |
| Entropy-RegBackbone=Qwen2.5-Math-7B2026.05 | 50.1 |