Multi-image hallucination evaluation on MUIRBench
62AccuracyQwen2.5-VL + CAPL
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5-VL + CAPLParams=7B2026.03 | 62 | |
| GLM4.1VBase + CAPLParams=9B2026.03 | 60.57 | |
| Qwen2.5-VLParams=7B2026.03 | 58.42 | |
| GLM4.1VBaseParams=9B2026.03 | 57.84 | |
| InternVL2.5 + CAPLParams=8B2026.03 | 52.12 | |
| InternVL2.5Params=8B2026.03 | 48.54 | |
| InternVL2Params=7B2026.03 | 45.61 | |
| Qwen2VLParams=7B2026.03 | 39.57 | |
| Idefics3Params=8B2026.03 | 30.96 | |
| LLaVA-OVParams=7B2026.03 | 30.85 | |
| LLaVA-NextParams=7B2026.03 | 30.5 | |
| Idefics2Params=8B2026.03 | 29.85 |