Hallucination Rate Measurement on Medical Diagnostic Hallucination Dataset S2 Single
95Auto Detection RateQwen2-VL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen2-VL2026.05 | 95 | 86.6 | |
| InternVL22026.05 | 85.1 | 81.3 | |
| BiomedCLIP2026.05 | 81.1 | 42.1 | |
| GPT-5.12026.05 | 78.8 | 55.2 | |
| Gemini 3 Flash2026.05 | 66.5 | 50.6 | |
| LLaVA-NeXT2026.05 | 66.5 | 71.9 |