Hallucination Rate Measurement on Medical Diagnostic Hallucination Dataset S1 Normal
89.1Auto Detection RateGemini 3 Flash
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Gemini 3 Flash2026.05 | 89.1 | 66.7 | |
| InternVL22026.05 | 70.8 | 73.2 | |
| LLaVA-NeXT2026.05 | 69.3 | 75.8 | |
| GPT-5.12026.05 | 54 | 42.9 | |
| Qwen2-VL2026.05 | 43.4 | 55.2 | |
| BiomedCLIP2026.05 | 7.6 | 32.8 |