Grounding on MMbench
91.3AccuracyMAI-UI-32B
Evaluation Results
| Method | Links | |
|---|---|---|
| MAI-UI-32Bdesign=end-to-end2026.02 | 91.3 | |
| MAI-UI-8Bdesign=end-to-end2026.02 | 88.8 | |
| UI-Venus-1.5-30B-A3Bdesign=end-to-end2026.02 | 88.6 | |
| UI-Venus-1.5-8Bdesign=end-to-end2026.02 | 88.1 | |
| Holo2-30B-A3Breproduction=true, design=end-to-end2026.02 | 86.8 | |
| UI-Venus-1.0-72Bdesign=end-to-end2026.02 | 86.3 | |
| Step-GUI-8Bdesign=end-to-end2026.02 | 85.6 | |
| Holo2-8Breproduction=true, design=end-to-end2026.02 | 84.5 | |
| Step-GUI-4Bdesign=end-to-end2026.02 | 84 | |
| Qwen3-VL-30B-A3Breproduction=true, design=end-to-end2026.02 | 83.7 | |
| GUI-Owl-32Bdesign=end-to-end2026.02 | 83 | |
| MAI-UI-2Bdesign=end-to-end2026.02 | 82.6 | |
| Qwen3-VL-8Breproduction=true, design=end-to-end2026.02 | 81.4 | |
| GUI-Owl-7Bdesign=end-to-end2026.02 | 80.5 | |
| UI-Venus-1.5-2Bdesign=end-to-end2026.02 | 80.3 | |
| UI-Venus-1.0-7Bdesign=end-to-end2026.02 | 79.9 | |
| Qwen3-VL-2Breproduction=true, design=end-to-end2026.02 | 69.5 | |
| UI-TARS-1.5-7Bdesign=end-to-end2026.02 | 64.3 |