Visual Perception and Grounding on V*
88.9AccuracyDeepEyes
Evaluation results
| Method | Links | |
|---|---|---|
| DeepEyesSize=7B2026.05 | 88.9 | |
| MoCASize=7B2026.05 | 86.6 | |
| Pixel ReasonerSize=7B2026.05 | 84.3 | |
| Qwen2.5-VL-InstructSize=72B2026.05 | 81.2 | |
| Llava-OVSize=7B2026.05 | 72.8 | |
| Qwen2.5-VL-InstructSize=7B2026.05 | 71.4 | |
| VL-RethinkerSize=7B2026.05 | 68.2 | |
| mPLUG-Owl3Size=7B2026.05 | 64.5 | |
| R1-VLSize=7B2026.05 | 60.3 | |
| GPT-4o-miniSize=-2026.05 | 50.8 | |
| GPT-4oSize=-2026.05 | 45 | |
| DocopilotSize=8B2026.05 | 40.1 |