Mathematical Problem Solving on Omni-MATH (Best-of-N accuracy)
35.4Best-of-N AccuracyMSV 64
Evaluation Results
| Method | Links | |
|---|---|---|
| MSV 64N=642026.03 | 35.4 | |
| ProbeN=642026.03 | 33.35 | |
| Probe + WVN=642026.03 | 32.1 | |
| MSV 16N=162026.03 | 31.98 | |
| MSV 1N=642026.03 | 31.79 | |
| MSV 1 + WVN=642026.03 | 31.79 | |
| ProbeN=162026.03 | 31.76 | |
| Probe + WVN=162026.03 | 31.62 | |
| MSV 1 + WVN=162026.03 | 31.33 | |
| MSV 1N=162026.03 | 31.02 | |
| Probe + WVN=42026.03 | 29.1 | |
| MSV 1 + WVN=42026.03 | 29 | |
| MSV 4N=42026.03 | 28.91 | |
| ProbeN=42026.03 | 28.85 | |
| MSV 1N=42026.03 | 28.78 | |
| ProbeN=12026.03 | 24.32 | |
| MSV 1N=12026.03 | 24.32 |