# Offline Reinforcement Learning on D4RL Gym-MuJoCo v2 (Medium, Medium-Replay, Medium-Expert) Source: [Offline Reinforcement Learning on D4RL Gym-MuJoCo v2 (Medium, Medium-Replay, Medium-Expert)]() Collection index: [llms.txt]() Data updated: 2026-05-12T07:20:16.140Z Evaluation of Offline Reinforcement Learning performance on the D4RL Gym-MuJoCo v2 environments, specifically reporting scores for the Medium, Medium-Replay, and Medium-Expert dataset configurations. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Offline Reinforcement Learning | | Dataset (including any stated split) | D4RL Gym-MuJoCo v2 | | Metric | HalfCheetah-Medium Return | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 13 of 13 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [HalfCheetah-Medium Return](): Higher is better; unit: SCALAR. - [Hopper-Medium Return](): Higher is better; unit: SCALAR. - [Walker2d-Medium Return](): Higher is better; unit: SCALAR. - [HalfCheetah-Medium-Replay Return](): Higher is better; unit: SCALAR. - [Hopper-Medium-Replay Return](): Higher is better; unit: SCALAR. - [Walker2d-Medium-Replay Return](): Higher is better; unit: SCALAR. - [HalfCheetah-Medium-Expert Return](): Higher is better; unit: SCALAR. - [Hopper-Medium-Expert Return](): Higher is better; unit: SCALAR. - [Walker2d-Medium-Expert Return](): Higher is better; unit: SCALAR. - [D4RL MuJoCo v2 Average Return](): Higher is better; unit: SCALAR. ## Recorded results | Method | HalfCheetah-Medium Return | Result date | Report paper | | --- | --- | --- | --- | | ACL-QL | 69.8 | Not recorded | 6a02b3ebdbc836b87d824520 | | A2PR | 68.6 | Not recorded | 6a02b3ebdbc836b87d824520 | | DOSER | 67.5 | Not recorded | 6a02b3ebdbc836b87d824520 | | SVR | 60.5 | Not recorded | 6a02b3ebdbc836b87d824520 | | SRPO | 60.4 | Not recorded | 6a02b3ebdbc836b87d824520 | | DTQL | 57.9 | Not recorded | 6a02b3ebdbc836b87d824520 | | QGPO | 54.1 | Not recorded | 6a02b3ebdbc836b87d824520 | | DQL | 51.5 | Not recorded | 6a02b3ebdbc836b87d824520 | | IDQL | 51 | Not recorded | 6a02b3ebdbc836b87d824520 | | TD3+BC | 48.3 | Not recorded | 6a02b3ebdbc836b87d824520 | | IQL | 47.4 | Not recorded | 6a02b3ebdbc836b87d824520 | | SfBC | 45.9 | Not recorded | 6a02b3ebdbc836b87d824520 | | CQL | 44 | Not recorded | 6a02b3ebdbc836b87d824520 | ### ACL-QL HalfCheetah-Medium Return: 69.8. Result ID: 6a02d46147f34ce924e53114. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method category":"Conventional methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction. ### A2PR HalfCheetah-Medium Return: 68.6. Result ID: 6a02d46147f34ce924e53111. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method category":"Conventional methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction. ### DOSER HalfCheetah-Medium Return: 67.5. Result ID: 6a02d46147f34ce924e5311b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method category":"Diffusion-based methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction. ### SVR HalfCheetah-Medium Return: 60.5. Result ID: 6a02d46147f34ce924e53113. [Reporting paper]() Recorded evaluation configuration: {"Method category":"Conventional methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction. ### SRPO HalfCheetah-Medium Return: 60.4. Result ID: 6a02d46147f34ce924e53119. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method category":"Diffusion-based methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction. ### DTQL HalfCheetah-Medium Return: 57.9. Result ID: 6a02d46147f34ce924e5311a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method category":"Diffusion-based methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction. ### QGPO HalfCheetah-Medium Return: 54.1. Result ID: 6a02d46147f34ce924e53118. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method category":"Diffusion-based methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction. ### DQL HalfCheetah-Medium Return: 51.5. Result ID: 6a02d46147f34ce924e53115. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method category":"Diffusion-based methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction. ### IDQL HalfCheetah-Medium Return: 51. Result ID: 6a02d46147f34ce924e53117. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method category":"Diffusion-based methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction. ### TD3+BC HalfCheetah-Medium Return: 48.3. Result ID: 6a02d46147f34ce924e5310f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Method category":"Conventional methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction. ### IQL HalfCheetah-Medium Return: 47.4. Result ID: 6a02d46147f34ce924e53110. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method category":"Conventional methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction. ### SfBC HalfCheetah-Medium Return: 45.9. Result ID: 6a02d46147f34ce924e53116. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method category":"Diffusion-based methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction. ### CQL HalfCheetah-Medium Return: 44. Result ID: 6a02d46147f34ce924e53112. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method category":"Conventional methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.