Visual Question Answering on Inhouse VQA
76.77Pass@1MedXIAOHE
Evaluation Results
| Method | Links | |
|---|---|---|
| MedXIAOHEMode=Thinking mode, Decoding=Greedy2026.02 | 76.77 | |
| Gemini 2.5 ProDecoding=Greedy2026.02 | 69.64 | |
| GPT-5.2 ThinkingMode=Thinking mode, Decoding=Greedy2026.02 | 68.27 |