Hallucination Rate Measurement on Medical Diagnostic Hallucination Dataset Multi 2-3 S3
92.8Auto Detection RateQwen2-VL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen2-VL2026.05 | 92.8 | 87.4 | |
| InternVL22026.05 | 76.8 | 65.9 | |
| GPT-5.12026.05 | 76.2 | 49.7 | |
| BiomedCLIP2026.05 | 75.5 | 37.2 | |
| Gemini 3 Flash2026.05 | 66.7 | 31 | |
| LLaVA-NeXT2026.05 | 66.7 | 73.6 |