Instruction Following on IFBench Out-of-Distribution (F1 Metrics)
41.5Initial F1AutoPyVerifier
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| AutoPyVerifierVerifier Model=Claude Hai-4.52026.04 | 41.5 | 63.1 | 21.6 | |
| AutoPyVerifierVerifier Model=Gemini-3.1 Fl2026.04 | 33.5 | 41.3 | 7.8 | |
| AutoPyVerifierVerifier Model=GPT-5.42026.04 | 25.1 | 37.1 | 12 | |
| AutoPyVerifierVerifier Model=Gemini-3.1 Pro2026.04 | 25.1 | 79.5 | 54.4 | |
| AutoPyVerifierVerifier Model=GPT-5.4 mini2026.04 | 22.4 | 41.5 | 19.1 |