# Multi-agent Reinforcement Learning on SMAC 25m v1 (test)

Source: [Multi-agent Reinforcement Learning on SMAC 25m v1 (test)](<https://www.sota2.com/research/sota/multi-agent-reinforcement-learning-on-smac-25m-v1-test>)

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

Data updated: 2026-05-19T10:12:10.411Z

Evaluation of Multi-Agent Reinforcement Learning performance on the test split of the SMAC 25m v1 scenario, typically measured by test win rate.

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 | Multi-agent Reinforcement Learning |
| Dataset (including any stated split) | SMAC 25m v1 (test) |
| Metric | Test Win Rate |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 9 of 9 |
| 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

- [Test Win Rate](<https://www.sota2.com/research/sota/multi-agent-reinforcement-learning-on-smac-25m-v1-test.md?metric=Test%20Win%20Rate>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Test Win Rate | Result date | Report paper |
| --- | --- | --- | --- |
| CommFormer | 98.3 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| QMIX | 98.1 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| MAGI | 97.5 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| HIBCG | 97.2 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| HIB-flat | 96.5 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| GACG | 96 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| BVME | 95.9 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| AIB-only | 94.3 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| ExpoComm | 91.5 | Not recorded | 6a0bee9ddbc836b87d82c492 |

### CommFormer

Test Win Rate: 98.3. Result ID: 6a0c36bafbc373592c02db5d.

[Reporting paper](<https://www.sota2.com/research/paper/6a0bee9ddbc836b87d82c492.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73351e209131cea5b287.md>)

Recorded evaluation configuration: {"Environment Type":"homo. control","Training Budget":"2M"}. This record does not assert independent reproduction.

### QMIX

Test Win Rate: 98.1. Result ID: 6a0c36bafbc373592c02db5b.

[Reporting paper](<https://www.sota2.com/research/paper/6a0bee9ddbc836b87d82c492.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64671e209131cea046e9.md>)

Recorded evaluation configuration: {"Environment Type":"homo. control","Training Budget":"2M"}. This record does not assert independent reproduction.

### MAGI

Test Win Rate: 97.5. Result ID: 6a0c36bafbc373592c02db5c.

[Reporting paper](<https://www.sota2.com/research/paper/6a0bee9ddbc836b87d82c492.md>)

Recorded evaluation configuration: {"Environment Type":"homo. control","Training Budget":"2M"}. This record does not assert independent reproduction.

### HIBCG

Test Win Rate: 97.2. Result ID: 6a0c36bafbc373592c02db63.

[Reporting paper](<https://www.sota2.com/research/paper/6a0bee9ddbc836b87d82c492.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0bee9ddbc836b87d82c492.md>)

Recorded evaluation configuration: {"Environment Type":"homo. control","Training Budget":"2M"}. This record does not assert independent reproduction.

### HIB-flat

Test Win Rate: 96.5. Result ID: 6a0c36bafbc373592c02db62.

[Reporting paper](<https://www.sota2.com/research/paper/6a0bee9ddbc836b87d82c492.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0bee9ddbc836b87d82c492.md>)

Recorded evaluation configuration: {"Environment Type":"homo. control","Training Budget":"2M"}. This record does not assert independent reproduction.

### GACG

Test Win Rate: 96. Result ID: 6a0c36bafbc373592c02db5f.

[Reporting paper](<https://www.sota2.com/research/paper/6a0bee9ddbc836b87d82c492.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73271e209131cea5ab8a.md>)

Recorded evaluation configuration: {"Environment Type":"homo. control","Training Budget":"2M"}. This record does not assert independent reproduction.

### BVME

Test Win Rate: 95.9. Result ID: 6a0c36bafbc373592c02db60.

[Reporting paper](<https://www.sota2.com/research/paper/6a0bee9ddbc836b87d82c492.md>)

Recorded evaluation configuration: {"Environment Type":"homo. control","Training Budget":"2M"}. This record does not assert independent reproduction.

### AIB-only

Test Win Rate: 94.3. Result ID: 6a0c36bafbc373592c02db61.

[Reporting paper](<https://www.sota2.com/research/paper/6a0bee9ddbc836b87d82c492.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0bee9ddbc836b87d82c492.md>)

Recorded evaluation configuration: {"Environment Type":"homo. control","Training Budget":"2M"}. This record does not assert independent reproduction.

### ExpoComm

Test Win Rate: 91.5. Result ID: 6a0c36bafbc373592c02db5e.

[Reporting paper](<https://www.sota2.com/research/paper/6a0bee9ddbc836b87d82c492.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa772e1e209131cea719f9.md>)

Recorded evaluation configuration: {"Environment Type":"homo. control","Training Budget":"2M"}. This record does not assert independent reproduction.

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