Binary equivalence judgment on French medical OEQA
76.19PrecisionGPT-5.1
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GPT-5.1Judge=GPT-5.12026.03 | 76.19 | 32 | 45.07 | 76.6 | 38.27 | |
| Gemini-2.5-ProJudge=Gemini-2.5-Pro2026.03 | 65.35 | 44 | 52.59 | 76.2 | 38.81 | |
| Qwen-80BJudge=Qwen-80B2026.03 | 64.62 | 56 | 60 | 77.6 | 44.77 | |
| MedGemma-27BJudge=MedGemma-27B2026.03 | 52.17 | 72 | 60.5 | 71.8 | 40.67 | |
| GRPO-Phi-3.5-miniJudge=GRPO-Phi-3.5-mini2026.03 | 51.91 | 63.33 | 57.06 | 71.4 | 36.33 | |
| SFT-Phi-3.5-miniJudge=SFT-Phi-3.5-mini2026.03 | 38.81 | 91.33 | 54.47 | 54.2 | 29.79 | |
| Phi-3.5-miniJudge=Phi-3.5-mini2026.03 | 35.94 | 98 | 52.59 | 47 | 27.49 |