Short-term Planning on AndroidControl Low
85.17TypeGUI-R1-7B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GUI-R1-7Bmode=Post-Training2026.01 | 85.17 | 84.02 | 66.52 | |
| SSL-7Bmode=Post-Training2026.01 | 85.17 | 87.08 | 70.72 | |
| QwenVL2.5-7B*mode=Post-Training, SFT=Mix-3K2026.01 | 84 | 85.74 | 64.32 | |
| GUI-R1-3Bmode=Post-Training2026.01 | 83.68 | 81.59 | 64.41 | |
| SSL-3Bmode=Post-Training2026.01 | 83.57 | 84.63 | 81.5 | |
| QwenVL2.5-7Bmode=Zero Shot2026.01 | 83.44 | 87.08 | 62.5 | |
| RL-Continous-7Bmode=Post-Training2026.01 | 83.12 | 83.34 | 65.81 | |
| RL-Continous-3Bmode=Post-Training2026.01 | 82.93 | 84.08 | 68.28 | |
| RL-Binary-7Bmode=Post-Training2026.01 | 81.87 | 81.59 | 64.75 | |
| RL-Binary-3Bmode=Post-Training2026.01 | 81.78 | 80.93 | 65.77 | |
| UI-R1-3Bmode=Post-Training2026.01 | 79.15 | 82.41 | 66.44 | |
| GPT-4omode=Zero Shot2026.01 | 74.33 | 38.67 | 28.39 | |
| Os-Atlas-7Bmode=Post-Training2026.01 | 73 | 73.37 | 50.94 | |
| QwenVL2.5-3B*mode=Post-Training, SFT=Mix-3K2026.01 | 71.08 | 74.53 | 58.79 | |
| Os-Atlas-4Bmode=Post-Training2026.01 | 64.58 | 71.19 | 40.62 | |
| QwenVL2.5-3Bmode=Zero Shot2026.01 | 62.03 | 74.07 | 59.32 |