Pairwise Preference Prediction
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
78.3Preference Rate (A > B)
17
May 26, 2026
65.6Preference Rate (A > B)
17
May 26, 2026
70.2P(A > B)
17
May 26, 2026
77.6Accuracy
9
Jun 23, 2026
85.6Accuracy
9
Jun 23, 2026
70.5Accuracy
9
Jun 23, 2026
79.1Aesthetic Score
7
Jun 15, 2026
69.8Alignment Score
7
Jun 15, 2026
72.2Accuracy
7
Jun 15, 2026
66.8Accuracy
7
Jun 15, 2026
91.61Accuracy
4
Jun 19, 2026
83.5Accuracy
4
Jun 19, 2026
70.4Accuracy
4
Jun 19, 2026
86.85Accuracy
4
Jun 19, 2026
90.37Accuracy
4
Jun 19, 2026
96.29Accuracy
4
Jun 19, 2026
74.72Accuracy
4
Jun 19, 2026
73.27Accuracy
4
Jun 19, 2026
83.84Accuracy
4
Jun 19, 2026
100Preference Accuracy (USC Franka)
4
Mar 4, 2026
90.11Accuracy
1
Jul 3, 2026
85.39Accuracy
1
Jul 3, 2026
78.54Accuracy
1
Jul 3, 2026
83.33Accuracy
1
Jul 3, 2026