Mathematical Problem Solving on OlymMATH-EN
48.3AccuracyKACE
Evaluation Results
| Method | Links | |
|---|---|---|
| KACESolver model=gpt-4.1-mini2026.05 | 48.3 | |
| Tiered Best-of-NSolver model=gpt-4.1-mini2026.05 | 44.4 | |
| Tiered + GEPASolver model=gpt-4.1-mini2026.05 | 44.4 | |
| Best-of-5 self-consistencySolver model=gpt-4.1-mini2026.05 | 40.7 | |
| Tiered + ACESolver model=gpt-4.1-mini2026.05 | 40.7 |