Autoformalization on ProverBench
95.38Success CountGoedel-V2-Formalizer-32B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Goedel-V2-Formalizer-32BPass@k=4, Computational budget=4 inference calls2026.04 | 95.38 | 83.38 | |
| DSRPass@k=4, Computational budget=4 inference calls2026.04 | 95.38 | 84 | |
| Goedel-V2-Formalizer-8BPass@k=4, Computational budget=4 inference calls2026.04 | 94.77 | 82.15 | |
| Kimina-Autoformalizer-7BPass@k=4, Computational budget=4 inference calls2026.04 | 93.23 | 65.54 | |
| Qwen3-MaxPass@k=4, Computational budget=4 inference calls2026.04 | 82.15 | 68.62 | |
| StepFun-Formalizer-32BPass@k=4, Computational budget=4 inference calls2026.04 | 78.77 | 66.77 | |
| StepFun-Formalizer-7BPass@k=4, Computational budget=4 inference calls2026.04 | 76.92 | 57.54 |