Medical VQA on MSD In-Domain (test)
83.08AccuracyVITAL
Evaluation Results
| Method | Links | |
|---|---|---|
| VITAL2026.05 | 83.08 | |
| Fleming-VL2026.05 | 65.5 | |
| GPT-5.42026.05 | 57.65 | |
| HuatuoGPT-Vision2026.05 | 55.57 | |
| Claude-Opus-4.62026.05 | 54.26 | |
| LVRWeights=Fine-tuned2026.05 | 48.63 | |
| SIM-CoTBackbone=Qwen3-VL-8B + LoRA2026.05 | 47.25 | |
| LLaVA-Med2026.05 | 46.74 | |
| Qwen3-VL-8BThinking=true2026.05 | 42.52 | |
| CODIBackbone=Qwen3-VL-8B + LoRA2026.05 | 41.09 | |
| CoconutBackbone=Qwen3-VL-8B + LoRA2026.05 | 40.58 | |
| VividMed2026.05 | 40.19 | |
| MCOUTWeights=Fine-tuned2026.05 | 38.54 | |
| Qwen3-VL-8BZero-shot=true2026.05 | 35.56 | |
| LVRWeights=Original2026.05 | 30.22 | |
| MCOUTWeights=Original2026.05 | 27.63 |