Visual Perception and Reasoning on V* Bench 1.0 (test)
83.48Attribute ScoreSR-MCR-7B
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| SR-MCR-7BScale=7B2025.12 | 83.48 | 76.32 | 80.63 | |
| WeThink-7BScale=7B2025.12 | 82.61 | 75 | 79.58 | |
| SR-MCR-3BScale=3B2025.12 | 81.16 | 61.6 | 73.12 | |
| Qwen2.5-VL-7BScale=7B2025.12 | 80.87 | 75 | 78.53 | |
| Qwen2.5-VL-3BScale=3B2025.12 | 78.26 | 60.53 | 71.2 | |
| InternVL3-8BScale=8B2025.12 | 70.43 | 72.37 | 71.2 | |
| Keye-VL-8B ThinkingScale=8B, Thinking mode=true2025.12 | 66.96 | 75 | 70.16 | |
| VL-Rethinker-7BScale=7B2025.12 | 64.35 | 71.05 | 67.02 | |
| GPT-4oAccess=Closed-source2025.12 | 62.03 | 72 | 66 | |
| InternVL3-2BScale=2B2025.12 | 59.13 | 63.16 | 60.73 | |
| VLAA-Thinker-7BScale=7B2025.12 | 56.52 | 59.21 | 57.59 | |
| Kimi-VL-A3B ThinkingScale=A3B, Thinking mode=true2025.12 | 51.3 | 60.53 | 54.97 | |
| SAIL-VL2-8B ThinkingScale=8B, Thinking mode=true2025.12 | 51.3 | 65.79 | 57.07 |