Formal Math Proving on MiniF2F Lean4 (test)
69.7Pass@16 (Overall)PhysProver
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| PhysProverSampling=Pass@16, Trajectories=16, Training Data=PhysLeanData2026.01 | 69.7 | 20 | 46.7 | 55.6 | 92.9 | 88.3 | 44.4 | 50 | 50 | |
| Deepseek-Prover-V2Sampling=Pass@16, Trajectories=162026.01 | 68.4 | 20 | 53.3 | 55.6 | 90 | 85 | 44.4 | 50 | 50 |