Mathematical Reasoning on MATH500 (avg@1)
93.6Average Score (avg@1)NPR
Evaluation Results
| Method | Links | |
|---|---|---|
| NPRData=orz-8k, Train=Parallel RL, Base=NPR-BETA2025.12 | 93.6 | |
| Qwen3-4B-Instruct-2507Data=N/A, Train=N/A, Base=N/A2025.12 | 93.4 | |
| SRData=orz-8k, Train=Sequential RL, Base=NPR-BETA2025.12 | 92.8 | |
| SR-BETAData=orz-8k, Train=Sequential SFT, Base=Q3-4B-Inst.2025.12 | 92 | |
| Multiverse-32BData=s1.1-8k, Train=S→P SFT, Base=Q2.5-32B-Inst.2025.12 | 91.8 | |
| NPR (Variant)Data=orz-8k, Train=Parallel RL, Base=NPR-BETA2025.12 | 91.8 | |
| NPR-BETAData=orz-8k, Train=Parallel SFT, Base=Q3-4B-Inst.2025.12 | 91.6 | |
| NPR-BETA (Variant)Data=orz-8k, Train=Parallel SFT, Base=Q3-4B2025.12 | 86.8 | |
| Qwen3-4B (Non-Thinking)Data=N/A, Train=N/A, Base=N/A2025.12 | 84.8 | |
| Multiverse-4BData=s1.1-8k, Train=S→P SFT, Base=Q3-4B-Inst.2025.12 | 81.6 | |
| Qwen2.5-32B-InstructData=N/A, Train=N/A, Base=N/A2025.12 | 80.4 |