Visual Grounding on V* Bench
95.7Overall Success Rateo3-0416
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| o3-0416Model Category=Private Models2025.11 | 95.7 | — | — | |
| GRiPModel Category=Open-source Visual Grounded Reasoning Models, Base Model=Qwen2.5-VL-7B2025.11 | 91.9 | 95.4 | 88.3 | |
| TreeVGR-7BModel Category=Open-source Visual Grounded Reasoning Models2025.11 | 91.1 | 94 | 87 | |
| DeepEyes-7BModel Category=Open-source Visual Grounded Reasoning Models2025.11 | 90 | 92.1 | 86.8 | |
| Qwen2.5-VL-32BModel Category=Open-source General Models2025.11 | 85.9 | 83.5 | 89.5 | |
| Qwen2.5-VL-72BModel Category=Open-source General Models2025.11 | 84.8 | 90.8 | 80.9 | |
| Pixel-Reasoner-7BModel Category=Open-source Visual Grounded Reasoning Models2025.11 | 80.6 | 83.5 | 76.3 | |
| InternVL3-38BModel Category=Open-source General Models2025.11 | 77.5 | 77.4 | 77.6 | |
| InternVL3-78BModel Category=Open-source General Models2025.11 | 76.4 | 75.7 | 77.6 | |
| ViGoRL-3bTraining data=UGround, Prompted to think=true2025.05 | 75.13 | — | — | |
| ViGoRL-3bTraining data=SAT-2, Prompted to think=true2025.05 | 74.87 | — | — | |
| Qwen2.5-VL-7BModel Category=Open-source General Models2025.11 | 74.3 | 77.4 | 69.7 | |
| Qwen2.5-VL-3B BasePrompted to think=false2025.05 | 74.21 | — | — | |
| LLaVA-OneVision-72BModel Category=Open-source General Models2025.11 | 73.8 | 80.9 | 63.2 | |
| InternVL3-8BModel Category=Open-source General Models2025.11 | 72.3 | 73 | 71.1 | |
| LLaVA-OneVision-7BModel Category=Open-source General Models2025.11 | 70.7 | 73 | 60.5 | |
| GPT-4o-1120Model Category=Private Models2025.11 | 66 | — | — |