Mathematical Reasoning on AIME Out-of-Distribution
69.1Initial F1AutoPyVerifier
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| AutoPyVerifierVerifier Model=GPT-5.4 mini2026.04 | 69.1 | 73.3 | 4.2 | |
| AutoPyVerifierVerifier Model=Gemini-3.1 Pro2026.04 | 66.1 | 77.3 | 11.2 | |
| AutoPyVerifierVerifier Model=GPT-5.42026.04 | 50.1 | 74.8 | 24.7 | |
| AutoPyVerifierVerifier Model=Gemini-3.1 Fl2026.04 | 41.2 | 91.6 | 50.4 | |
| AutoPyVerifierVerifier Model=Claude Hai-4.52026.04 | 34.1 | 53.3 | 19.2 |