Mathematical Theorem Proving on PutnamBench (Cost Metrics)
163Total SpendHilbert
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| HilbertBase Model=Gemini 2.5 Pro, Budget (pass@k)=pass@12026.06 | 163 | 244 | — | — | |
| Goedel-ArchitectBase Model=DeepSeek-V4-Flash2026.06 | 294 | 0.44 | 0.21 | 1.14 | |
| AxProverBase Model=AxProverBase, Budget (pass@k)=pass@12026.06 | 8,467 | 12.6 | — | — |