Cross-domain Reasoning on ComplexFuncBench Out-of-Distribution
49.7Initial F1AutoPyVerifier
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| AutoPyVerifierVerifier Model=GPT-5.42026.04 | 49.7 | 49.7 | 0 | |
| AutoPyVerifierVerifier Model=GPT-5.4 mini2026.04 | 49.7 | 46.1 | -3.6 | |
| AutoPyVerifierVerifier Model=Gemini-3.1 Fl2026.04 | 49.7 | 55.3 | 5.6 | |
| AutoPyVerifierVerifier Model=Claude Hai-4.52026.04 | 49.7 | 49.7 | 0 | |
| AutoPyVerifierVerifier Model=Gemini-3.1 Pro2026.04 | 23.7 | 74.3 | 50.6 |