Mathematical Problem Solving on Olympiad (Pass@1)
42.13Pass@1VHG (Soft)
Evaluation results
| Method | Links | |
|---|---|---|
| VHG (Soft)2026.05 | 42.13 | |
| Vanilla-GRPOtraining_variant=w. SFT data2026.05 | 39.91 | |
| Vanilla-GRPO2026.05 | 39.6 | |
| R-ZeroIteration=12026.05 | 36.16 | |
| R-ZeroIteration=22026.05 | 35.41 | |
| Qwen3-4B-Base2026.05 | 34.88 | |
| R-ZeroIteration=32026.05 | 34.06 |