Agent Task Performance
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
13.57Attack Success Rate
24
Feb 26, 2026
62.5Attack Success Rate
18
Feb 26, 2026
79.4Retail Performance
16
Jun 5, 2026
63.4Overall Performance
11
Jun 16, 2026
64.5Overall Success Rate
11
Jun 16, 2026
69.79Utility
6
Feb 26, 2026
67.62Utility
6
Feb 26, 2026
66.67Utility
6
Feb 26, 2026
64.4Success Rate (1-)
4
Jun 2, 2026
66.7Success Rate (1-step)
4
Jun 2, 2026
81.2G Score
2
Jul 8, 2026
54.7G Score
2
Jul 8, 2026
0.519G
2
Jul 8, 2026