Cross-domain Reasoning on ComplexFuncBench In-Distribution
55.6Initial F1AutoPyVerifier
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| AutoPyVerifierVerifier Model=Gemini-3.1 Pro2026.04 | 55.6 | 96.4 | 40.8 | |
| AutoPyVerifierVerifier Model=Gemini-3.1 Fl2026.04 | 47.9 | 84.9 | 37 | |
| AutoPyVerifierVerifier Model=GPT-5.42026.04 | 30.1 | 85.1 | 55 | |
| AutoPyVerifierVerifier Model=GPT-5.4 mini2026.04 | 27.7 | 62.4 | 34.7 | |
| AutoPyVerifierVerifier Model=Claude Hai-4.52026.04 | 27.7 | 54 | 26.3 |