Multimodal Relation Reasoning on R-Bench
85.05AccuracyChainMPQ
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ChainMPQModel=InternVL3.52025.10 | 85.05 | 82.85 | 85.91 | |
| PromptingModel=InternVL3.52025.10 | 83.97 | 80.31 | 84.87 | |
| ChainMPQModel=Qwen2.5-VL2025.10 | 83.92 | 80.23 | 84.63 | |
| CoTModel=InternVL3.52025.10 | 82.87 | 79.41 | 84.02 | |
| CalibrateModel=InternVL3.52025.10 | 82.65 | 79.82 | 83.75 | |
| VanillaModel=InternVL3.52025.10 | 82.33 | 78.87 | 83.21 | |
| PromptingModel=Qwen2.5-VL2025.10 | 81.02 | 78.38 | 82.46 | |
| CalibrateModel=Qwen2.5-VL2025.10 | 80.98 | 77.05 | 81.52 | |
| CoTModel=Qwen2.5-VL2025.10 | 80.24 | 76.94 | 81.64 | |
| VanillaModel=Qwen2.5-VL2025.10 | 79.85 | 75.19 | 80.88 | |
| ChainMPQModel=LLaVA-1.52025.10 | 76.04 | 72.03 | 81.54 | |
| PromptingModel=LLaVA-1.52025.10 | 75.86 | 67.28 | 79.6 | |
| ChainMPQModel=InstructBLIP2025.10 | 75.86 | 70.59 | 81.12 | |
| CalibrateModel=LLaVA-1.52025.10 | 74.28 | 67.86 | 78.31 | |
| PromptingModel=InstructBLIP2025.10 | 73.65 | 68.74 | 80.21 | |
| CalibrateModel=InstructBLIP2025.10 | 72.96 | 66.23 | 78.54 | |
| CoTModel=LLaVA-1.52025.10 | 72.91 | 65.88 | 77.87 | |
| CoTModel=InstructBLIP2025.10 | 72.15 | 65.79 | 78.45 | |
| VanillaModel=LLaVA-1.52025.10 | 71.23 | 64.27 | 77.28 | |
| VanillaModel=InstructBLIP2025.10 | 69.31 | 62.76 | 76.04 |