Mathematical Reasoning on AIME 2024 (Acc, Cost)
40AccuracyCOPE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| COPEShot=Zero, Train=×2025.06 | 40 | 478 | |
| LargeShot=Zero, Train=×2025.06 | 33.3 | 502 | |
| ABCShot=Zero, Train=×2025.06 | 33.3 | 485 | |
| FrugalGPTShot=Few, Train=✓2025.06 | 33.3 | 513 | |
| Cascade (verify)Shot=Few, Train=×2025.06 | 26.6 | 570 | |
| Cascade (vote)Shot=Few, Train=×2025.06 | 23.3 | 568 | |
| Small (majority)Shot=Zero, Train=×2025.06 | 6.67 | — | |
| Small (single)Shot=Zero, Train=×2025.06 | 3.33 | — |