Mathematical Reasoning on MATH500 (0-shot)
47.4Pass@1HEX
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| HEXGen. Len.=512, Training Necessity=Training-free2026.02 | 47.4 | 6,400 | |
| DVOTINGGen. Len.=512, Training Necessity=Training-free2026.02 | 45.4 | 701.2 | |
| Majority VotingGen. Len.=512, Training Necessity=Training-free2026.02 | 44 | 1,280 | |
| HEXGen. Len.=256, Training Necessity=Training-free2026.02 | 43.6 | 3,200 | |
| IGPOGen. Len.=256, Training Necessity=RL2026.02 | 42.8 | 256 | |
| d1Gen. Len.=512, Training Necessity=SFT + RL2026.02 | 40.2 | 256 | |
| DVOTINGGen. Len.=256, Training Necessity=Training-free2026.02 | 40.2 | 473.7 | |
| HEXGen. Len.=128, Training Necessity=Training-free2026.02 | 39.6 | 1,600 | |
| RFGGen. Len.=256, Training Necessity=FT or RL2026.02 | 39.6 | 512 | |
| wd1Gen. Len.=512, Training Necessity=RL2026.02 | 39 | 256 | |
| d1Gen. Len.=256, Training Necessity=SFT + RL2026.02 | 38.6 | 128 | |
| LLaDA-8B-InstructGen. Len.=512, Training Necessity=Pre-training2026.02 | 38.2 | 512 | |
| Majority VotingGen. Len.=256, Training Necessity=Training-free2026.02 | 36.8 | 640 | |
| LLaDA-8B-InstructGen. Len.=256, Training Necessity=Pre-training2026.02 | 35 | 256 | |
| DVOTINGGen. Len.=128, Training Necessity=Training-free2026.02 | 34.8 | 292.1 | |
| wd1Gen. Len.=256, Training Necessity=RL2026.02 | 34.4 | 128 | |
| d1Gen. Len.=128, Training Necessity=SFT + RL2026.02 | 33.8 | 64 | |
| Majority VotingGen. Len.=128, Training Necessity=Training-free2026.02 | 33 | 320 | |
| LLaDA-8B-InstructGen. Len.=128, Training Necessity=Pre-training2026.02 | 31.4 | 128 |