# Reinforcement Learning on Gymnasium HalfCheetah

Source: [Reinforcement Learning on Gymnasium HalfCheetah](<https://www.sota2.com/research/sota/reinforcement-learning-on-gymnasium-halfcheetah>)

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

Data updated: 2026-03-16T08:59:07.499Z

Evaluation of Reinforcement Learning performance on the Gymnasium HalfCheetah continuous control environment, typically measured by Return.

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 | Reinforcement Learning |
| Dataset (including any stated split) | Gymnasium HalfCheetah |
| Metric | Return |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [Return](<https://www.sota2.com/research/sota/reinforcement-learning-on-gymnasium-halfcheetah.md?metric=Return>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Return | Result date | Report paper |
| --- | --- | --- | --- |
| CAPO | 2737 | Not recorded | 69b78e590636e5bf4afbda08 |
| CAPO-Avg | 1967 | Not recorded | 69b78e590636e5bf4afbda08 |
| Best-of-K | 1940 | Not recorded | 69b78e590636e5bf4afbda08 |
| TRPO | 1629 | Not recorded | 69b78e590636e5bf4afbda08 |
| PPO | 1604 | Not recorded | 69b78e590636e5bf4afbda08 |
| PPO-K× | 1532 | Not recorded | 69b78e590636e5bf4afbda08 |
| PPO-SWA | 1261 | Not recorded | 69b78e590636e5bf4afbda08 |

### CAPO

Return: 2737. Result ID: 69b7c56147f34ce9249f8c79.

[Reporting paper](<https://www.sota2.com/research/paper/69b78e590636e5bf4afbda08.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b78e590636e5bf4afbda08.md>)

Recorded evaluation configuration: {"Backbone":"(64, 64) MLP","Training steps":"1M","Seeds":"8","Compute-matched":"true"}. This record does not assert independent reproduction.

### CAPO-Avg

Return: 1967. Result ID: 69b7c56147f34ce9249f8c78.

[Reporting paper](<https://www.sota2.com/research/paper/69b78e590636e5bf4afbda08.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b78e590636e5bf4afbda08.md>)

Recorded evaluation configuration: {"Backbone":"(64, 64) MLP","Training steps":"1M","Seeds":"8","Compute-matched":"true"}. This record does not assert independent reproduction.

### Best-of-K

Return: 1940. Result ID: 69b7c56147f34ce9249f8c75.

[Reporting paper](<https://www.sota2.com/research/paper/69b78e590636e5bf4afbda08.md>)

Recorded evaluation configuration: {"Backbone":"(64, 64) MLP","Training steps":"1M","Seeds":"8","Compute-matched":"true"}. This record does not assert independent reproduction.

### TRPO

Return: 1629. Result ID: 69b7c56147f34ce9249f8c77.

[Reporting paper](<https://www.sota2.com/research/paper/69b78e590636e5bf4afbda08.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa634f1e209131ce9fe056.md>)

Recorded evaluation configuration: {"Backbone":"(64, 64) MLP","Training steps":"1M","Seeds":"8","Compute-matched":"false"}. This record does not assert independent reproduction.

### PPO

Return: 1604. Result ID: 69b7c56147f34ce9249f8c73.

[Reporting paper](<https://www.sota2.com/research/paper/69b78e590636e5bf4afbda08.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63a01e209131ce9ffdaa.md>)

Recorded evaluation configuration: {"Backbone":"(64, 64) MLP","Training steps":"1M","Seeds":"8","Compute-matched":"false"}. This record does not assert independent reproduction.

### PPO-K×

Return: 1532. Result ID: 69b7c56147f34ce9249f8c74.

[Reporting paper](<https://www.sota2.com/research/paper/69b78e590636e5bf4afbda08.md>)

Recorded evaluation configuration: {"Backbone":"(64, 64) MLP","Training steps":"1M","Seeds":"8","Compute-matched":"true"}. This record does not assert independent reproduction.

### PPO-SWA

Return: 1261. Result ID: 69b7c56147f34ce9249f8c76.

[Reporting paper](<https://www.sota2.com/research/paper/69b78e590636e5bf4afbda08.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65651e209131cea09f96.md>)

[Implementation](<https://github.com/timgaripov/swa>)

Recorded evaluation configuration: {"Backbone":"(64, 64) MLP","Training steps":"1M","Seeds":"8","Compute-matched":"false"}. This record does not assert independent reproduction.

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