GUI Localization on GUIZoom-Bench hard-mislead
54.1Localization AccuracyQwen3-VL-32B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-VL-32BZoom Depth=d42025.12 | 54.1 | |
| Qwen3-VL-8BZoom Depth=d42025.12 | 45.9 | |
| Qwen3-VL-32BZoom Depth=d22025.12 | 45.9 | |
| Qwen3-VL-32BZoom Depth=d32025.12 | 45.9 | |
| Qwen3-VL-8BZoom Depth=d32025.12 | 40.5 | |
| UI-Venus-7BZoom Depth=d22025.12 | 35.1 | |
| UI-Venus-7BZoom Depth=d32025.12 | 35.1 | |
| Qwen3-VL-8BZoom Depth=d22025.12 | 32.4 | |
| UI-Venus-7BZoom Depth=d42025.12 | 29.7 | |
| UI-TARS-7BZoom Depth=d32025.12 | 24.3 | |
| UI-TARS-1.5-7BZoom Depth=d22025.12 | 24.3 | |
| Qwen3-VL-32BZoom Depth=d12025.12 | 21.6 | |
| UI-TARS-1.5-7BZoom Depth=d32025.12 | 21.6 | |
| UI-TARS-7BZoom Depth=d42025.12 | 18.9 | |
| UI-TARS-1.5-7BZoom Depth=d12025.12 | 18.9 | |
| UI-TARS-1.5-7BZoom Depth=d42025.12 | 18.9 | |
| UI-Venus-7BZoom Depth=d12025.12 | 18.9 | |
| Qwen3-VL-8BZoom Depth=d12025.12 | 16.2 | |
| Qwen2.5-VL-7BZoom Depth=d32025.12 | 16.2 | |
| Qwen2.5-VL-7BZoom Depth=d42025.12 | 13.5 | |
| Qwen2.5-VL-7BZoom Depth=d22025.12 | 10.8 | |
| UI-TARS-7BZoom Depth=d22025.12 | 10.8 | |
| Claude-4.5-SonnetZoom Depth=d12025.12 | 10.5 | |
| GPT-5Zoom Depth=d22025.12 | 8.1 | |
| GPT-5Zoom Depth=d42025.12 | 5.4 | |
| UI-TARS-7BZoom Depth=d12025.12 | 5.4 | |
| GPT-5Zoom Depth=d32025.12 | 2.7 | |
| GPT-5Zoom Depth=d12025.12 | 0 | |
| Qwen2.5-VL-7BZoom Depth=d12025.12 | 0 |