Offline Reinforcement Learning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
94.3Normalized Return
10
May 29, 2026
96.4D4RL Normalized Score
10
May 29, 2026
37Test Return
10
May 15, 2026
67.9Test Return
10
May 15, 2026
25Normalized Score
10
Apr 27, 2026
43Normalized Score
10
Apr 27, 2026
74.5Normalized Performance
10
Mar 4, 2026
59Success Rate
10
Mar 4, 2026
14Success Rate
10
Mar 4, 2026
87.4Normalized Score
10
Mar 18, 2026
96Return (HalfCheetah, M-E)
10
Feb 26, 2026
3,967BP->DC*
10
Feb 26, 2026
96.8umaze Return
10
Feb 26, 2026
87.5Umaze Score
10
Feb 26, 2026
77.84Reward
10
Feb 26, 2026
85AntMaze UMaze (Fixed) Score
10
Feb 26, 2026
187.8Score
10
Feb 26, 2026
152.3Score
10
Feb 26, 2026
111Overall Score
10
Feb 26, 2026
60Scene Score
9
Jun 23, 2026
68.2HalfCheetah Medium v2 Score
9
Jun 9, 2026
91.5Normalized Score
9
May 19, 2026
59Average Task Success
9
May 12, 2026
41Average Task Success Rate
9
May 12, 2026
72Average Task Success
9
May 12, 2026