Task Success Rate on MobileMiniWob++
80.4Task Success RateGUI-Explorer
Evaluation results
| Method | Links | |
|---|---|---|
| GUI-ExplorerType=GPT-4o, Input=SoM2026.05 | 80.4 | |
| EAMType=GPT-4o, Qwen2.5-3B-instruct-ft, Input=SoM2026.05 | 76.1 | |
| AppAgentXType=GPT-4o, Input=SoM2026.05 | 72.8 | |
| M3AType=GPT-4o, Input=SoM2026.05 | 68.5 | |
| Qwen2.5-VL-72B2025.02 | 68 | |
| SoMInput=Image + AXTree, Planner=GPT-4-Turbo2024.12 | 67.7 | |
| Aguvis-72B2025.02 | 66 | |
| GPT-4o2025.02 | 61 | |
| ClaudeSet-of-Mark (SoM)=true2025.02 | 61 | |
| Aria-UIInput=Image, Planner=GPT-4o2024.12 | 60.4 | |
| ChoiceInput=AXTree, Planner=GPT-4-Turbo2024.12 | 59.7 | |
| ChoiceInput=AXTree, Planner=Gemini 1.5 Pro2024.12 | 57.4 | |
| GPT-4oType=GPT-4o, Input=SoM2026.05 | 56.5 | |
| UI-TARS-7BType=UI-TARS-7B, Input=screen2026.05 | 53.3 | |
| AutoDroid-V2Type=Llama-3-8B-ft, Input=SoM2026.05 | 53.3 | |
| Qwen2-VL-72BSet-of-Mark (SoM)=true2025.02 | 50 | |
| UGroundInput=Image, Planner=GPT-4o2024.12 | 48.4 | |
| Gemini 2.0Set-of-Mark (SoM)=true2025.02 | 42 | |
| SoMInput=Image + AXTree, Planner=Gemini 1.5 Pro2024.12 | 40.3 | |
| Qwen 2.5-VL-3BType=Qwen 2.5-VL-3B, Input=SoM2026.05 | 32.6 | |
| UI-TARS-2BType=UI-TARS-2B, Input=screen2026.05 | 31.5 |