Mathematical Problem Solving on HMMT 2025 (Pass@k)
64.44Pass@12OPSD
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| OPSDModel=Qwen3-4B, Max Generation Length=38,912 tokens, Number of Evaluation Seeds=32026.05 | 64.44 | 43.89 | 51.11 | |
| EOPDModel=Qwen3-4B, Max Generation Length=38,912 tokens, Number of Evaluation Seeds=32026.05 | 63.33 | 41.94 | 50 | |
| PW-OPSD ModerateModel=Qwen3-4B, Max Generation Length=38,912 tokens, Number of Evaluation Seeds=3, Schedule Type=Moderate2026.05 | 60 | 43.33 | 52.22 | |
| PW-OPSD AggressiveModel=Qwen3-4B, Max Generation Length=38,912 tokens, Number of Evaluation Seeds=3, Schedule Type=Aggressive2026.05 | 60 | 45.37 | 52.22 | |
| REOPOLDModel=Qwen3-4B, Max Generation Length=38,912 tokens, Number of Evaluation Seeds=32026.05 | 58.89 | 41.39 | 51.11 |