GUI Action Prediction and Grounding on OmniACT Web
89.22Type RateGUICrafter-3B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GUICrafter-3BEvaluation Protocol=Fine-Tuning, Training Stage=Stage 1 + Stage 22026.06 | 89.22 | 77.21 | 77.37 | |
| GUICrafter-3BEvaluation Protocol=Fine-Tuning, Training Stage=only Stage12026.06 | 88.63 | 65.85 | 66.23 | |
| GUI-R1-3BEvaluation Protocol=Fine-Tuning2026.06 | 88.58 | 75.1 | 75.08 | |
| OS-Atlas-7BEvaluation Protocol=Fine-Tuning2026.06 | 85.63 | 69.35 | 59.15 | |
| GUICrafter-3BEvaluation Protocol=Fine-Tuning, Training Stage=only Stage22026.06 | 85.35 | 75.33 | 75.33 | |
| Qwen2.5-VL-7B*Evaluation Protocol=Fine-Tuning, SFT dataset=GUI-R1-3K2026.06 | 81.62 | 73.45 | 73.39 | |
| GPT-4oEvaluation Protocol=Zero Shot2026.06 | 79.33 | 42.79 | 34.06 | |
| Qwen2.5-VL-7BEvaluation Protocol=Zero Shot2026.06 | 79.15 | 71.32 | 71.21 | |
| UI-R1-3BEvaluation Protocol=Fine-Tuning2026.06 | 75.42 | 61.35 | 61.33 | |
| Qwen2.5-VL-3B*Evaluation Protocol=Fine-Tuning, SFT dataset=GUI-R1-3K2026.06 | 66.24 | 56.91 | 53 | |
| Qwen2.5-VL-3BEvaluation Protocol=Zero Shot2026.06 | 50.63 | 46.89 | 47.02 | |
| OS-Atlas-4BEvaluation Protocol=Fine-Tuning2026.06 | 46.74 | 49.24 | 22.99 |