Multimodal Benchmarking on MMB 1.1
82.2AccuracyGPT-4o
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-4o2025.09 | 82.2 | |
| DeFacto (Ours)Backbone=Qwen2.5-VL-7B2025.09 | 81.2 | |
| Pixel Reasoner2025.09 | 78.5 | |
| Visual-SR12025.09 | 77.4 | |
| Chain-of-Focus2025.09 | 75.3 | |
| DeepEyes2025.09 | 29.4 |