Mathematical Reasoning on AIME (Best-of-K Accuracy)
41.3Best-of-K AccuracyVeriBound
Evaluation Results
| Method | Links | |
|---|---|---|
| VeriBoundK=5, Evaluation Mode=Empirical2026.06 | 41.3 | |
| FOVERK=52026.06 | 40.7 | |
| VeriBoundK=5, Evaluation Mode=Bound2026.06 | 39.5 | |
| ReasonFlux-PRMK=52026.06 | 39.1 | |
| R-PRMK=52026.06 | 38.5 | |
| PRM800KK=52026.06 | 37.8 | |
| Math-ShepherdK=52026.06 | 34.1 | |
| ORMK=52026.06 | 28.4 |