Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Offline Reinforcement Learning research benchmarks – Page 36 · SOTA2 Research

Offline Reinforcement Learning

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

VD4RL Cheetah-run pixel-based (expert)

46.8Normalized Score

3

Feb 26, 2026

VD4RL Cheetah-run pixel-based (medium-expert)

61Normalized Score

3

Feb 26, 2026

VD4RL Cheetah-run pixel-based (medium)

53.3Normalized Score

3

Feb 26, 2026

D4RL Mujoco Halfcheetah-Medium v2

42.3Normalized Score

3

Feb 26, 2026

Theoretical Complexity Analysis

1Sequential NFE

2

Jun 2, 2026

visual-cube-double-play singletask task1 v0

27Normalized Score

2

May 29, 2026

visual-cube single-play singletask task1 v0

89Normalized Score

2

May 29, 2026

MIMIC-III Sepsis (test)

68.68P-C Agreement

2

Feb 26, 2026

MIMIC-III Sepsis

0.496Vaso Dev.

2

Feb 26, 2026

D4RL Medium-Expert v2

90.7HalfCheetah Normalized Score

2

Feb 26, 2026

D4RL Medium-Replay v2

44.6HalfCheetah Score

2

Feb 26, 2026

48.3Normalized Score (HalfCheetah)

2

Feb 26, 2026

D4RL AntMaze Large-Diverse v0

0.2Avg Normalized Score

1

Feb 26, 2026

D4RL AntMaze Medium-Diverse v0

10.6Normalized Score

1

Feb 26, 2026

96.7HalfCheetah Normalized Score

1

Feb 26, 2026

D4RL v2 (Random)

11HalfCheetah Score

1

Feb 26, 2026

Theoretical Analysis

—

—Primary metric

0

Feb 26, 2026

V-D4RL Pixel-Based

—

—Primary metric

0

Feb 26, 2026

D4RL Antmaze-umaze-diverse 10k (reduced)

—

—Primary metric

0

Feb 26, 2026

D4RL Antmaze-umaze 10k (reduced)

—

—Primary metric

0

Feb 26, 2026

Page 36 of 36

Popular datasets

D4RL Walker2D halfcheetah OGBench hopper

Follow for updates

Join SOTA2 Contact