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

MemGUI-Bench

Benchmarks

Task NameDataset NameSOTA ResultTrend
GUI Agent TaskMemGUI-Bench (Overall)
Precision@137.5
19
GUI Agent TaskMemGUI-Bench Hard
P@134.2
19
GUI Agent TaskMemGUI-Bench Medium
Precision@135.7
19
GUI Agent TaskMemGUI-Bench Easy
P@141.7
19
Long-horizon GUI InteractionMemGUI-Bench
Precision@1 (1 App)50
14
Task SuccessMemGUI-Bench 1.0 (test)
Pass@1 L166.7
11
Mobile GUI Agent NavigationMemGUI-Bench
Success Rate42.97
4
Showing 7 of 7 rows