Mobile Agent Safety and Capability Evaluation on Phone-Harm + Normal-150 (merged)
0.2Harm Rate (HR)GPT-5
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GPT-5Setting=Base2026.04 | 0.2 | 68.75 | 4.26 | 0.29 | 0.4 | |
| Gemini-3Setting=Base2026.04 | 0.42 | 71.67 | 10.53 | 0.59 | 3.85 | |
| AutoGLM-VLMSetting=Base2026.04 | 1.79 | 72.22 | 39.56 | 2.48 | 0.79 | |
| CORA2026.04 | 2.19 | 89.69 | 85.29 | 2.44 | 9.83 | |
| UI-TARS-1.5Setting=Base2026.04 | 2.87 | 77.28 | 57.14 | 3.71 | 3.85 |