# Offline Reinforcement Learning on Antmaze v0 (test)

Source: [Offline Reinforcement Learning on Antmaze v0 (test)](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-antmaze-v0-test>)

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

Data updated: 2026-05-26T07:38:06.390Z

Evaluation of Offline Reinforcement Learning performance across the full Antmaze v0 suite (umaze, medium, large, play, diverse configurations) on the test set.

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) | Antmaze v0 (test) |
| Metric | Success Rate (medium play) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| 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

- [Success Rate (medium play)](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-antmaze-v0-test.md?metric=Success%20Rate%20(medium%20play)>): Higher is better; unit: PERCENT.

- [Success Rate (umaze)](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-antmaze-v0-test.md?metric=Success%20Rate%20(umaze)>): Higher is better; unit: PERCENT.

- [Success Rate (umaze diverse)](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-antmaze-v0-test.md?metric=Success%20Rate%20(umaze%20diverse)>): Higher is better; unit: PERCENT.

- [Success Rate (medium diverse)](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-antmaze-v0-test.md?metric=Success%20Rate%20(medium%20diverse)>): Higher is better; unit: PERCENT.

- [Success Rate (large play)](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-antmaze-v0-test.md?metric=Success%20Rate%20(large%20play)>): Higher is better; unit: PERCENT.

- [Success Rate (large diverse)](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-antmaze-v0-test.md?metric=Success%20Rate%20(large%20diverse)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Success Rate (medium play) | Result date | Report paper |
| --- | --- | --- | --- |
| TT+Q | 93.3 | Not recorded | 6a1528e1dbd7150506f7dc83 |
| TAP+G | 78 | Not recorded | 6a1528e1dbd7150506f7dc83 |
| LEQ | 58.8 | Not recorded | 6a1528e1dbd7150506f7dc83 |
| RAMBO | 16.4 | Not recorded | 6a1528e1dbd7150506f7dc83 |
| COMBO | 0 | Not recorded | 6a1528e1dbd7150506f7dc83 |

### TT+Q

Success Rate (medium play): 93.3. Result ID: 6a154db5fbc373592c0c5c59.

[Reporting paper](<https://www.sota2.com/research/paper/6a1528e1dbd7150506f7dc83.md>)

Recorded evaluation configuration: {"Variant":"Q-guided beam search"}. This record does not assert independent reproduction.

### TAP+G

Success Rate (medium play): 78. Result ID: 6a154db5fbc373592c0c5c5a.

[Reporting paper](<https://www.sota2.com/research/paper/6a1528e1dbd7150506f7dc83.md>)

Recorded evaluation configuration: {"Variant":"goal-conditioned"}. This record does not assert independent reproduction.

### LEQ

Success Rate (medium play): 58.8. Result ID: 6a154db5fbc373592c0c5c5b.

[Reporting paper](<https://www.sota2.com/research/paper/6a1528e1dbd7150506f7dc83.md>)

Recorded evaluation configuration: {"Evaluation protocol":"Averaged over 5 seeds"}. This record does not assert independent reproduction.

### RAMBO

Success Rate (medium play): 16.4. Result ID: 6a154db5fbc373592c0c5c58.

[Reporting paper](<https://www.sota2.com/research/paper/6a1528e1dbd7150506f7dc83.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6da51e209131cea3b264.md>)

[Implementation](<https://www.sota2.com/github.com/marc-rigter/rambo>)

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

### COMBO

Success Rate (medium play): 0. Result ID: 6a154db5fbc373592c0c5c57.

[Reporting paper](<https://www.sota2.com/research/paper/6a1528e1dbd7150506f7dc83.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6da51e209131cea3b264.md>)

[Implementation](<https://www.sota2.com/github.com/marc-rigter/rambo>)

Recorded evaluation configuration: {"Source":"Reported in RAMBO \[S4\]"}. This record does not assert independent reproduction.

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