GUI Automation Replay on AppAgent-Claw Workflow Benchmark baseline
50E2E Success RateAppAgent-Claw
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| AppAgent-ClawWorkflow=Overall, Machine setting=intended same machine, Baseline conditions=true2026.04 | 50 | 100 | 20.5 | 85.2 | 7.1 | 7.6 | |
| AppAgent-ClawWorkflow=Notes: create new note, Steps/run=10, Clicks/run=4, Machine setting=intended same machine, Baseline conditions=true2026.04 | 10 | 100 | 19.7 | 75 | 0 | 25 | |
| AppAgent-ClawWorkflow=Safari: open search tab, Steps/run=7, Clicks/run=2, Machine setting=intended same machine, Baseline conditions=true2026.04 | 10 | 100 | 11.3 | 100 | 0 | 0 | |
| AppAgent-ClawWorkflow=NetEase Music: play daily recommendation, Steps/run=11, Clicks/run=6, Machine setting=intended same machine, Baseline conditions=true2026.04 | 10 | 100 | 28.7 | 83.3 | 16.7 | 0 | |
| AppAgent-ClawWorkflow=NetEase Music: play favorite songs, Steps/run=11, Clicks/run=6, Machine setting=intended same machine, Baseline conditions=true2026.04 | 10 | 100 | 24.7 | 100 | 0 | 0 | |
| AppAgent-ClawWorkflow=WeChat: send message to File Transfer Assistant, Steps/run=13, Clicks/run=3, Machine setting=intended same machine, Baseline conditions=true2026.04 | 10 | 100 | 18.1 | 63.3 | 16.7 | 20 |