Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

GUI-Act-Web

Benchmarks

Task NameDataset NameSOTA ResultTrend
Short-term planningGUI-Act-Web
Type Success Rate94.54
16
GUI Interaction ControlGUI-Act-Web
Type Accuracy96.3
10
GUI Grounding and ActionGUI-Act-Web (OOD)
Type Acc89.36
8
GUI Agent NavigationGUI-Act-Web (test)
Type Accuracy89.9
4
GUI reasoningGUI-Act-Web
Type Success Rate87.73
4
Showing 5 of 5 rows