Medical Question Answering on Medbullets op5 (Pass@1)
94.16Pass@1MedXIAOHE
Evaluation Results
| Method | Links | |
|---|---|---|
| MedXIAOHEMode=Thinking mode, Decoding=Greedy2026.02 | 94.16 | |
| GPT-5.2 ThinkingMode=Thinking mode, Decoding=Greedy2026.02 | 88.31 | |
| Gemini 3.0 ProDecoding=Greedy2026.02 | 86.04 | |
| Gemini 2.5 ProDecoding=Greedy2026.02 | 72.73 |