Hallucination Rate Measurement on Medical Diagnostic Hallucination Dataset Complex >=4 S4
85.3Auto Detection RateQwen2-VL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen2-VL2026.05 | 85.3 | 83 | |
| InternVL22026.05 | 72.7 | 67 | |
| GPT-5.12026.05 | 70 | 39.5 | |
| Gemini 3 Flash2026.05 | 68.7 | 33.5 | |
| LLaVA-NeXT2026.05 | 66.8 | 77.6 | |
| BiomedCLIP2026.05 | 65.7 | 33 |