Mathematical Reasoning on MATH-500 (Pass@1 Accuracy and Avg.)
80Pass@1 AccuracyScaf-GRPO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Scaf-GRPOCategory=Long CoT2025.10 | 80 | 50.9 | |
| rStar-Math-7BCategory=Long CoT2025.10 | 78.4 | 43.6 | |
| Oat-ZeroCategory=Long CoT2025.10 | 78 | 46.5 | |
| OpenReasoner-ZeroCategory=Long CoT2025.10 | 77.4 | 45 | |
| SimpleRL-ZeroCategory=Long CoT2025.10 | 76.8 | 42.6 | |
| LUFFYCategory=Long CoT2025.10 | 75.2 | 46.6 | |
| Eurus-2-7B-PRIMECategory=Long CoT2025.10 | 74.4 | 43.6 | |
| Qwen-2.5-7B-BaseCategory=Short CoT2025.10 | 61.8 | 31.3 | |
| Qwen-2.5-Math-7B-BaseCategory=Short CoT2025.10 | 53.6 | 27.5 |