Reward Modeling
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
78.9Accuracy
40
Feb 26, 2026
82Overall
40
Feb 26, 2026
84.7General
39
May 12, 2026
0.145MAE
38
May 8, 2026
63Accuracy
36
May 13, 2026
88.8Accuracy
36
May 13, 2026
0.074MAE
36
May 8, 2026
80.7Accuracy
36
Apr 8, 2026
83.3Accuracy
34
Feb 26, 2026
92.42Target Adherence
32
Jun 4, 2026
0.709Average Accuracy
32
Jun 4, 2026
89.5Accuracy
30
Mar 5, 2026
74.3Overall Score
29
Apr 14, 2026
95.5Accuracy
28
Mar 4, 2026
65.46Prediction Accuracy
28
Feb 26, 2026
69.01Prediction Accuracy
28
Feb 26, 2026
76.44Prediction Accuracy
28
Feb 26, 2026
93.6Final Token Accuracy
27
Apr 28, 2026
-0.629LE Score
26
Feb 26, 2026
-0.18EF
26
Feb 26, 2026
75PPE Corr
26
Feb 26, 2026
0.933RWBench
25
Jul 2, 2026
85.4PF
25
Jun 26, 2026
60.2Accuracy
24
May 13, 2026
74.9Accuracy
24
May 13, 2026