# Offline Reinforcement Learning on D4RL Gym Locomotion v2 (All Conditions)

Source: [Offline Reinforcement Learning on D4RL Gym Locomotion v2 (All Conditions)](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-05T08:30:16.155Z

Evaluation of Offline Reinforcement Learning performance across the HalfCheetah, Hopper, and Walker2D environments using the D4RL Gym Locomotion v2 dataset, aggregating results from Medium, Medium Replay, Medium Expert, and Random data distributions, along with a total locomotion score.

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 locomotion v2 (various) |
| Metric | HalfCheetah Medium Score |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 16 of 16 |
| 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 Score](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions.md?metric=HalfCheetah%20Medium%20Score>): Higher is better; unit: SCALAR.

- [Hopper Medium Score](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions.md?metric=Hopper%20Medium%20Score>): Higher is better; unit: SCALAR.

- [Walker2D Medium Score](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions.md?metric=Walker2D%20Medium%20Score>): Higher is better; unit: SCALAR.

- [HalfCheetah Medium Replay Score](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions.md?metric=HalfCheetah%20Medium%20Replay%20Score>): Higher is better; unit: SCALAR.

- [Hopper Medium Replay Score](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions.md?metric=Hopper%20Medium%20Replay%20Score>): Higher is better; unit: SCALAR.

- [Walker2D Medium Replay Score](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions.md?metric=Walker2D%20Medium%20Replay%20Score>): Higher is better; unit: SCALAR.

- [HalfCheetah Medium Expert Score](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions.md?metric=HalfCheetah%20Medium%20Expert%20Score>): Higher is better; unit: SCALAR.

- [Hopper Medium Expert Score](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions.md?metric=Hopper%20Medium%20Expert%20Score>): Higher is better; unit: SCALAR.

- [Walker2D Medium Expert Score](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions.md?metric=Walker2D%20Medium%20Expert%20Score>): Higher is better; unit: SCALAR.

- [HalfCheetah Random Score](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions.md?metric=HalfCheetah%20Random%20Score>): Higher is better; unit: SCALAR.

- [Hopper Random Score](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions.md?metric=Hopper%20Random%20Score>): Higher is better; unit: SCALAR.

- [Walker2D Random Score](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions.md?metric=Walker2D%20Random%20Score>): Higher is better; unit: SCALAR.

- [Locomotion Total Score](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-gym-locomotion-v2-all-conditions.md?metric=Locomotion%20Total%20Score>): Higher is better; unit: SCALAR.

## Recorded results

| Method | HalfCheetah Medium Score | Result date | Report paper |
| --- | --- | --- | --- |
| MOPO | 73.1 | Not recorded | 68fa75941e209131cea68bf0 |
| OneStep | 50.4 | Not recorded | 68fa75941e209131cea68bf0 |
| TD3BC | 48.3 | Not recorded | 68fa75941e209131cea68bf0 |
| IQL | 47.4 | Not recorded | 68fa75941e209131cea68bf0 |
| CQL | 47 | Not recorded | 68fa75941e209131cea68bf0 |
| TT | 46.9 | Not recorded | 68fa7bee1e209131cea8b5fb |
| SCAS | 46.6 | Not recorded | 68fa75941e209131cea68bf0 |
| SDC | 45.9 | Not recorded | 68fa75941e209131cea68bf0 |
| OSR | 45.1 | Not recorded | 68fa75941e209131cea68bf0 |
| CQL | 44.4 | Not recorded | 68fa7bee1e209131cea8b5fb |
| RATE | 43.5 | Not recorded | 68fa7bee1e209131cea8b5fb |
| ELMUR | 43.2 | Not recorded | 68fa7bee1e209131cea8b5fb |
| BC | 43.1 | Not recorded | 68fa7bee1e209131cea8b5fb |
| ELMUR | 43.1 | Not recorded | 68fa7bee1e209131cea8b5fb |
| DT | 42.6 | Not recorded | 68fa7bee1e209131cea8b5fb |
| BC | 42 | Not recorded | 68fa75941e209131cea68bf0 |

### MOPO

HalfCheetah Medium Score: 73.1. Result ID: 6993ef783a6c314849199bf7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75941e209131cea68bf0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68c11e209131cea1e111.md>)

Recorded evaluation configuration: {"runtime":"900m","hyperparameter tuning":"w/"}. This record does not assert independent reproduction.

### OneStep

HalfCheetah Medium Score: 50.4. Result ID: 6993ef783a6c314849199bf8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75941e209131cea68bf0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b631e209131cea2df51.md>)

[Implementation](<https://github.com/davidbrandfonbrener/onestep-rl>)

Recorded evaluation configuration: {"runtime":"120m","hyperparameter tuning":"w/o"}. This record does not assert independent reproduction.

### TD3BC

HalfCheetah Medium Score: 48.3. Result ID: 6993ef783a6c314849199bf9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75941e209131cea68bf0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b631e209131cea2df4f.md>)

[Implementation](<https://github.com/sfujim/TD3_BC>)

Recorded evaluation configuration: {"runtime":"60m","hyperparameter tuning":"w/o"}. This record does not assert independent reproduction.

### IQL

HalfCheetah Medium Score: 47.4. Result ID: 6993ef783a6c314849199bfb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75941e209131cea68bf0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6af71e209131cea2b67f.md>)

Recorded evaluation configuration: {"runtime":"100m","hyperparameter tuning":"w/o"}. This record does not assert independent reproduction.

### CQL

HalfCheetah Medium Score: 47. Result ID: 6993ef783a6c314849199bfa.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75941e209131cea68bf0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa682a1e209131cea1a6c2.md>)

Recorded evaluation configuration: {"runtime":"250m","hyperparameter tuning":"w/"}. This record does not assert independent reproduction.

### TT

HalfCheetah Medium Score: 46.9. Result ID: 69a93f056dbb7075b952a93c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bee1e209131cea8b5fb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b511e209131cea2d9cb.md>)

[Implementation](<https://www.sota2.com/github.com/jannerm/trajectory-transformer>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### SCAS

HalfCheetah Medium Score: 46.6. Result ID: 6993ef783a6c314849199bfe.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75941e209131cea68bf0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75941e209131cea68bf0.md>)

[Implementation](<https://github.com/maoyixiu/SCAS>)

Recorded evaluation configuration: {"runtime":"140m","hyperparameter tuning":"w/o"}. This record does not assert independent reproduction.

### SDC

HalfCheetah Medium Score: 45.9. Result ID: 6993ef783a6c314849199bfd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75941e209131cea68bf0.md>)

Recorded evaluation configuration: {"runtime":"420m","hyperparameter tuning":"w/"}. This record does not assert independent reproduction.

### OSR

HalfCheetah Medium Score: 45.1. Result ID: 6993ef783a6c314849199bfc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75941e209131cea68bf0.md>)

Recorded evaluation configuration: {"runtime":"300m","hyperparameter tuning":"w/"}. This record does not assert independent reproduction.

### CQL

HalfCheetah Medium Score: 44.4. Result ID: 69a93f056dbb7075b952a939.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bee1e209131cea8b5fb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa682a1e209131cea1a6c2.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### RATE

HalfCheetah Medium Score: 43.5. Result ID: 69a93f056dbb7075b952a93d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bee1e209131cea8b5fb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75361e209131cea66793.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### ELMUR

HalfCheetah Medium Score: 43.2. Result ID: 69a93f056dbb7075b952a93e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bee1e209131cea8b5fb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bee1e209131cea8b5fb.md>)

[Implementation](<https://github.com/CognitiveAISystems/RATE>)

Recorded evaluation configuration: {"Conditioning Mode":"(R, o, a) returns-to-go, observations, and actions"}. This record does not assert independent reproduction.

### BC

HalfCheetah Medium Score: 43.1. Result ID: 69a93f056dbb7075b952a93b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bee1e209131cea8b5fb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a381e209131cea26b46.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### ELMUR

HalfCheetah Medium Score: 43.1. Result ID: 69a93f056dbb7075b952a93f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bee1e209131cea8b5fb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bee1e209131cea8b5fb.md>)

[Implementation](<https://github.com/CognitiveAISystems/RATE>)

Recorded evaluation configuration: {"Conditioning Mode":"(o) observations alone"}. This record does not assert independent reproduction.

### DT

HalfCheetah Medium Score: 42.6. Result ID: 69a93f056dbb7075b952a93a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bee1e209131cea8b5fb.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a381e209131cea26b46.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### BC

HalfCheetah Medium Score: 42. Result ID: 6993ef783a6c314849199bf6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75941e209131cea68bf0.md>)

Recorded evaluation configuration: {"runtime":"30m","hyperparameter tuning":"w/o"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
