Binary disease diagnosis on RSNA OOD
79Macro AccuracyInternVL3.5-38B-Instruct
Evaluation results
| Method | Links | ||
|---|---|---|---|
| InternVL3.5-38B-Instruct2026.05 | 79 | 62.8 | |
| Hulu-Med-Flash-Preview-27B2026.05 | 79 | 29.1 | |
| CSRL2026.05 | 77.2 | 63 | |
| Qwen2.5-VL-7B-Instruct2026.05 | 77.1 | 33.6 | |
| Majority Voting (tie → neg)2026.05 | 74.5 | 60.2 | |
| MedGemma-27B-IT2026.05 | 74.2 | 58.5 | |
| Gemini-3.1-Pro-Preview2026.05 | 73.1 | 60.2 | |
| HuatuoGPT-Vision-7B2026.05 | 72.6 | 58 | |
| MedKLIP2026.05 | 71.4 | 58.9 | |
| Lingshu-7B2026.05 | 70.6 | 57.5 | |
| DenseNet-MIMIC2026.05 | 70.3 | 46.1 | |
| DenseNet-NIH2026.05 | 68.9 | 49.8 | |
| GPT-4o2026.05 | 68.7 | 55.6 | |
| DenseNet-CheX2026.05 | 68.6 | 50.5 | |
| CheXzero2026.05 | 68.3 | 56.3 | |
| Majority Voting (tie → pos)2026.05 | 68.2 | 56.9 | |
| CheXagent-8B2026.05 | 49.2 | 45.1 | |
| Best Tool per Pathology2026.05 | 49.2 | 45.1 | |
| Claude-Sonnet-4.62026.05 | 46.5 | 43.3 | |
| Logistic Regression2026.05 | 45.9 | 44.6 | |
| Qwen2.5-VL-72B-Instruct2026.05 | 42.5 | 42.9 |