Reward Modeling
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
80.5Accuracy
16
Feb 26, 2026
87.2Accuracy
16
Feb 26, 2026
74.4Accuracy
16
Feb 26, 2026
86Accuracy
16
Feb 26, 2026
93Accuracy (Wikipedia)
15
Jul 2, 2026
85Score
15
Jun 4, 2026
50.7BoN Score
15
Apr 14, 2026
67.7Accuracy
15
Apr 10, 2026
79Pairwise Accuracy
15
Mar 26, 2026
89.4Chat-Hard Score
15
Mar 18, 2026
0.985R2@50
14
Jun 19, 2026
0.794R2 (Top-50)
14
Jun 19, 2026
0.543R2@50
14
Jun 19, 2026
99.4Agreement Rate
14
Apr 8, 2026
62.3Accuracy
14
Apr 2, 2026
91.01Accuracy
14
Mar 12, 2026
92.5Precision
14
Feb 26, 2026
89.83Accuracy
13
Apr 15, 2026
86.25Accuracy
13
Apr 15, 2026
72.18Accuracy
13
Apr 15, 2026
86.45Accuracy
13
Apr 15, 2026
88.6Help Accuracy
13
Mar 10, 2026
76.4Accuracy
13
Mar 4, 2026
91.4Accuracy
13
May 8, 2026
89.19Pairwise Acc
13
Feb 26, 2026