Mathematical Reasoning on AIME 25 (avg@8)
53.8Avg@8 ScoreNPR (Variant)
Evaluation Results
| Method | Links | |
|---|---|---|
| NPR (Variant)Data=orz-8k, Train=Parallel RL, Base=NPR-BETA2025.12 | 53.8 | |
| NPRData=orz-8k, Train=Parallel RL, Base=NPR-BETA2025.12 | 50.4 | |
| SRData=orz-8k, Train=Sequential RL, Base=NPR-BETA2025.12 | 49.2 | |
| Qwen3-4B-Instruct-2507Data=N/A, Train=N/A, Base=N/A2025.12 | 47.4 | |
| Multiverse-32BData=s1.1-8k, Train=S→P SFT, Base=Q2.5-32B-Inst.2025.12 | 45.8 | |
| NPR-BETA (Variant)Data=orz-8k, Train=Parallel SFT, Base=Q3-4B2025.12 | 43.8 | |
| Multiverse-4BData=s1.1-8k, Train=S→P SFT, Base=Q3-4B-Inst.2025.12 | 42.9 | |
| NPR-BETAData=orz-8k, Train=Parallel SFT, Base=Q3-4B-Inst.2025.12 | 42.9 | |
| SR-BETAData=orz-8k, Train=Sequential SFT, Base=Q3-4B-Inst.2025.12 | 37.1 | |
| Qwen3-4B (Non-Thinking)Data=N/A, Train=N/A, Base=N/A2025.12 | 19.1 | |
| Qwen2.5-32B-InstructData=N/A, Train=N/A, Base=N/A2025.12 | 10.4 |