Offline GUI Agent Evaluation on CAGUI (Full)
90.3Type AccuracyMimo-VL-7B + WildGUI
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Mimo-VL-7B + WildGUIPre-training=WildGUI2026.05 | 90.3 | 71 | |
| UI-TARS-7BModel Category=Open-source Models2026.05 | 88.6 | 70.3 | |
| Qwen2.5-VL-7B* + WildGUIPre-training=WildGUI2026.05 | 88.3 | 65.4 | |
| Mimo-VL-7BModel Category=Open-source Models2026.05 | 82.2 | 63.4 | |
| OS-Atlas-7BModel Category=Open-source Models2026.05 | 81.5 | 55.9 | |
| Qwen2.5-VL-7B*Model Category=Open-source Models2026.05 | 74.2 | 55.2 | |
| Aguvis-7BModel Category=Open-source Models2026.05 | 67.4 | 38.2 | |
| OS-Genesis-7BModel Category=Open-source Models2026.05 | 38.1 | 14.5 | |
| GPT-4oModel Category=Closed-source Models2026.05 | 3.7 | 3.7 |