Human preference prediction
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
72.18Accuracy
51
Feb 26, 2026
85.4Accuracy
25
May 26, 2026
78.3Accuracy
21
May 12, 2026
82.2Accuracy
14
May 1, 2026
89.3Gov. Preference Score
12
May 19, 2026
66.7Accuracy
8
Mar 10, 2026
71.5Accuracy
8
Mar 10, 2026
74.24Overall
7
Feb 26, 2026
70.5Accuracy
7
Feb 26, 2026
58.46Accuracy (w/ Ties)
6
May 1, 2026
86.09Accuracy
6
Mar 31, 2026
67.75OPA
5
May 4, 2026
77.4Accuracy
4
Feb 26, 2026
39.62Recall @1
4
Feb 26, 2026
65.14Preference Accuracy
4
Feb 26, 2026