Instruction Following on IFBench In-Distribution (F1 Metrics)
41.1Initial F1AutoPyVerifier
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| AutoPyVerifierVerifier Model=Gemini-3.1 Fl2026.04 | 41.1 | 51.7 | 10.6 | |
| AutoPyVerifierVerifier Model=Claude Hai-4.52026.04 | 39.5 | 59.1 | 19.6 | |
| AutoPyVerifierVerifier Model=Gemini-3.1 Pro2026.04 | 29.4 | 81.5 | 52.1 | |
| AutoPyVerifierVerifier Model=GPT-5.42026.04 | 28.5 | 42.6 | 14.1 | |
| AutoPyVerifierVerifier Model=GPT-5.4 mini2026.04 | 25.7 | 39.5 | 13.8 |