Mathematical Theorem Proving on LEAN-IMO-BENCH (Advanced)
100Algebra AccuracyLEAP
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| LEAProllout=22026.06 | 100 | 25 | 100 | 12.5 | 56.7 | |
| Aristotlerollout=22026.06 | 37.5 | 12.5 | 33.3 | 0 | 20 | |
| Hilbertrollout=22026.06 | 12.5 | 0 | 16.6 | 0 | 6.6 | |
| Gemini-3.1-Propass@k=1282026.06 | 0 | 12.5 | 0 | 0 | 3.3 | |
| Goedel-V2-32Bpass@k=1282026.06 | 0 | 0 | 0 | 0 | 0 |