Mathematical Reasoning on MATH (Accuracy and Performance Gain)
82.98AccuracyQwen3-8B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen3-8BBase Model=Qwen3-8B2026.03 | 82.98 | — | |
| TraDOBase Model=Qwen3-8B2026.03 | 78.74 | 4.24 | |
| SDARBase Model=Qwen3-8B2026.03 | 74.22 | 8.76 | |
| Qwen2.5-7BBase Model=Qwen2.5-7B2026.03 | 73.06 | — | |
| dTRPOBase Model=Qwen2.5-7B2026.03 | 64.3 | 8.76 |