Mobile GUI interaction on AndroidLab
49.5Average SR (%)GUI-Owl-7B w/ Android Coach
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GUI-Owl-7B w/ Android Coach#Params=7B2026.04 | 49.5 | 58.5 | 45.2 | |
| GUI-Owl-7B w/ PPO#Params=7B2026.04 | 48.1 | 53.3 | 45.5 | |
| GUI-Owl-7B w/ GRPO#Params=7B2026.04 | 47.1 | 54.1 | 43.7 | |
| GUI-Owl-7B Base Model#Params=7B2026.04 | 44.9 | 48.1 | 43.4 | |
| UI-Genie-Agent#Params=72B2026.04 | 41.2 | — | — | |
| Claude-Sonnet-4 (SoM)#Params=-2026.04 | 40.6 | — | — | |
| UI-TARS-1.5-7B w/ Android Coach#Params=7B2026.04 | 39.4 | 42.6 | 33.7 | |
| UI-TARS-72B-DPO#Params=72B2026.04 | 35.5 | 36.4 | 31.5 | |
| UI-TARS-1.5-7B w/ PPO#Params=7B2026.04 | 35 | 38.9 | 29.4 | |
| UI-TARS-1.5-7B w/ GRPO#Params=7B2026.04 | 34.8 | 36.4 | 30.5 | |
| UI-TARS-1.5-7B Base Model#Params=7B2026.04 | 31.9 | 34 | 27.6 | |
| GPT-4o (SoM)#Params=-2026.04 | 31.2 | — | — | |
| Qwen2.5VL-32B-Instruct#Params=32B2026.04 | 28.5 | 28.4 | 25.8 | |
| Gemini-Pro-1.5 (SoM)#Params=-2026.04 | 16.7 | — | — | |
| AgentCPM-GUI-8B#Params=8B2026.04 | 14.7 | 8.6 | 16.8 | |
| Qwen2.5-VL-7B-Instruct#Params=7B2026.04 | 8.9 | 14.8 | 4.3 | |
| OS-Genesis-7B-AW#Params=7B2026.04 | 5.1 | 6.8 | 3.6 |