Object Hallucination Evaluation on GQA (Random)
90.13AccuracyCAI
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CAIBackbone=Qwen-VL2026.06 | 90.13 | 89.87 | |
| MESAModel=LLaVA-v1.52026.04 | 89.5 | 89.45 | |
| HDDModel=LLaVA-NeXT2025.12 | 89.3 | 89.4 | |
| ICTModel=LLaVA-v1.52026.04 | 89.26 | 88.78 | |
| CAIBackbone=LLaVA-1.52026.06 | 89.1 | 89.15 | |
| VAFModel=LLaVA-v1.52026.04 | 89 | 88.63 | |
| RITUALModel=LLaVA-NeXT2025.12 | 88.7 | 88.9 | |
| VTIModel=LLaVA-v1.52026.04 | 88.37 | 88.02 | |
| VCDBackbone=Qwen-VL2026.06 | 88.33 | 88.26 | |
| ONLYBackbone=Qwen-VL2026.06 | 88.13 | 87.54 | |
| CAIBackbone=InstructBLIP2026.06 | 88.1 | 87.84 | |
| MESAModel=Qwen-VL2026.04 | 87.97 | 86.67 | |
| NulluModel=LLaVA-v1.52026.04 | 87.7 | 88 | |
| HDDModel=LLaVA-1.52025.12 | 87.2 | 87.6 | |
| RITUALModel=LLaVA-1.52025.12 | 86.8 | 87.2 | |
| VCDModel=LLaVA-v1.52026.04 | 86.65 | 86.99 | |
| ICDModel=Qwen-VL2026.04 | 86.2 | 84.62 | |
| ONLYBackbone=LLaVA-1.52026.06 | 86.2 | 87.2 | |
| VAFModel=Qwen-VL2026.04 | 86.11 | 86.11 | |
| VCDModel=LLaVA-1.52025.12 | 86.1 | 87.1 | |
| ICDModel=LLaVA-v1.52026.04 | 85.9 | 85.29 | |
| PAIBackbone=Qwen-VL2026.06 | 85.9 | 84.77 | |
| VCDModel=LLaVA-NeXT2025.12 | 85.7 | 85.9 | |
| VCDModel=Qwen-VL2026.04 | 85.59 | 85.33 | |
| ICTModel=Qwen-VL2026.04 | 85.33 | 86.12 | |
| VanillaModel=LLaVA-v1.52026.04 | 84.96 | 86.25 | |
| VTIModel=Qwen-VL2026.04 | 84.33 | 81.94 | |
| HDDModel=InstructBLIP2025.12 | 84.1 | 84.9 | |
| ONLYBackbone=InstructBLIP2026.06 | 83.87 | 85.02 | |
| RegularBackbone=Qwen-VL2026.06 | 83.83 | 82.6 | |
| SamplingModel=LLaVA-1.52025.12 | 83.5 | 82.7 | |
| RITUALModel=InstructBLIP2025.12 | 83.5 | 84.3 | |
| PAIBackbone=LLaVA-1.52026.06 | 83.37 | 85.03 | |
| VCDBackbone=InstructBLIP2026.06 | 83.37 | 83.78 | |
| SamplingModel=LLaVA-NeXT2025.12 | 83.1 | 81.2 | |
| VCDModel=InstructBLIP2025.12 | 82.8 | 83.2 | |
| RegularBackbone=LLaVA-1.52026.06 | 82.03 | 83.86 | |
| VCDBackbone=LLaVA-1.52026.06 | 81.7 | 83.99 | |
| PAIBackbone=InstructBLIP2026.06 | 81.23 | 82.67 | |
| RegularBackbone=InstructBLIP2026.06 | 80.17 | 81.55 | |
| SamplingModel=InstructBLIP2025.12 | 79.5 | 80.4 | |
| VanillaModel=Qwen-VL2026.04 | 74.76 | 66.6 | |
| NulluModel=Qwen-VL2026.04 | 74.73 | 65.55 |