Computer Use
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
75OS Success Rate
45
Jun 2, 2026
75Score
12
May 20, 2026
27.8Clean Rate
10
May 26, 2026
68.3Success Rate
8
May 20, 2026
42.5Success Rate
6
Feb 26, 2026
27.6Writer Success Rate
5
Jun 18, 2026
50Success Rate (LibreOffice)
5
Jun 18, 2026
33.8Accuracy
4
May 13, 2026
42.1Success Rate
4
Feb 26, 2026
47.5Headline SR
3
May 26, 2026
34.5Headline SR
2
May 26, 2026
54Headline SR
1
May 26, 2026
89.2Headline Success Rate
1
May 26, 2026
72.6Headline SR
1
May 26, 2026
18.2Headline Success Rate
1
May 26, 2026
61.6Headline SR
1
May 26, 2026
28.7Headline SR
1
May 26, 2026
40.86Headline SR
1
May 26, 2026
51.1Headline Success Rate
1
May 26, 2026
59.1Headline SR
1
May 26, 2026
—
—Primary metric
0
May 26, 2026