Mathematical Problem Solving on MATH-HARD (Accuracy)
77.3Accuracy (MATH-HARD)KACE
Evaluation Results
| Method | Links | |
|---|---|---|
| KACESolver model=gpt-4o-mini2026.05 | 77.3 | |
| Tiered Best-of-NSolver model=gpt-4o-mini2026.05 | 73.3 | |
| Best-of-5 self-consistencySolver model=gpt-4o-mini2026.05 | 72.7 | |
| Tiered + GEPASolver model=gpt-4o-mini2026.05 | 72 | |
| Tiered + ACESolver model=gpt-4o-mini2026.05 | 70 |