Formal Theorem Proving on PutnamBench September 2025
462Solved Problems CountHILBERT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| HILBERTllm=Gemini 2.5 Pro, prover=Goedel-Prover-V2-32B2025.09 | 462 | 70 | |
| SeedProver2025.09 | 331 | 50.4 | |
| HILBERTllm=gpt-oss-120b, prover=Goedel-Prover-V2-32B2025.09 | 88 | 13.3 | |
| Goedel-Prover-V2-32Bself-correction=true, pass@k=1842025.09 | 86 | 13.4 | |
| DeepSeek-Prover-V2 671Bpass@k=10242025.09 | 47 | 7.1 | |
| Bourbakipass@k=5122025.09 | 26 | 4 | |
| DSP+pass@k=1282025.09 | 23 | 3.6 | |
| Kimina-Prover-7B-Distillpass@k=1922025.09 | 10 | 1.5 | |
| Self-play Theorem Proverpass@k=32002025.09 | 8 | 1.2 | |
| Goedel-Prover-SFTpass@k=5122025.09 | 7 | 1.1 | |
| ABELpass@k=5962025.09 | 7 | 1.1 |