GUI Grounding on MMBench-GUI L2 v1.0 (test)
88.8AccuracyMAI-UI
Evaluation Results
| Method | Links | |
|---|---|---|
| MAI-UIModel Scale=≈8B2026.06 | 88.8 | |
| VISTAModel Scale=≥30B, Inference Method=Multi-view (+ MVP)2026.06 | 87.6 | |
| VISTAModel Scale=≈8B, Inference Method=Multi-view (+ MVP)2026.06 | 87.3 | |
| VISTAModel Scale=≈4B, Inference Method=Multi-view (+ MVP)2026.06 | 86.8 | |
| VISTAModel Scale=≈8B, Inference Method=Single-view2026.06 | 86.8 | |
| Holo2Model Scale=≥30B2026.06 | 86.8 | |
| VISTAModel Scale=≥30B, Inference Method=Single-view2026.06 | 86.8 | |
| VISTAModel Scale=≈4B, Inference Method=Single-view2026.06 | 86.7 | |
| UI-VenusModel Scale=72B2026.06 | 86.3 | |
| Step-GUIModel Scale=≈8B2026.06 | 85.6 | |
| Qwen3-VLModel Scale=≈4B2026.06 | 85.3 | |
| Holo2Model Scale=≈8B2026.06 | 84.5 | |
| Step-GUIModel Scale=≈4B2026.06 | 84 | |
| Qwen3-VLModel Scale=≥30B2026.06 | 83.7 | |
| GUI-OwlModel Scale=≥30B2026.06 | 83 | |
| Qwen3-VLModel Scale=≈8B2026.06 | 81.3 | |
| GUI-OwlModel Scale=≈8B2026.06 | 80.5 | |
| UI-VenusModel Scale=≈8B2026.06 | 79.9 | |
| UI-TARS-1.5Model Scale=≈8B2026.06 | 64.3 |