# Multi-agent Battle on MAgent Battle n=36 agents (test)

Source: [Multi-agent Battle on MAgent Battle n=36 agents (test)](<https://www.sota2.com/research/sota/multi-agent-battle-on-magent-battle-n-36-agents-test>)

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

Data updated: 2026-05-19T10:12:10.450Z

Evaluation of multi-agent battle performance in the MAgent environment using 36 agents, measured by final return, peak return, and convergence 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 Battle |
| Dataset (including any stated split) | MAgent Battle n=36 agents 3M training steps (test) |
| Metric | Final Return |
| Unit | SCALAR |
| 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

- [Final Return](<https://www.sota2.com/research/sota/multi-agent-battle-on-magent-battle-n-36-agents-test.md?metric=Final%20Return>): Higher is better; unit: SCALAR.

- [Peak Return](<https://www.sota2.com/research/sota/multi-agent-battle-on-magent-battle-n-36-agents-test.md?metric=Peak%20Return>): Higher is better; unit: SCALAR.

- [Convergence Rate](<https://www.sota2.com/research/sota/multi-agent-battle-on-magent-battle-n-36-agents-test.md?metric=Convergence%20Rate>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Final Return | Result date | Report paper |
| --- | --- | --- | --- |
| HIBCG | 2.74 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| CommFormer | 2.5 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| QMIX | 2.19 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| AIB-only | 1.87 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| HIB-flat | 1.72 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| ExpoComm | 1.66 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| GACG | 1.63 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| BVME | 1.43 | Not recorded | 6a0bee9ddbc836b87d82c492 |
| MAGI | 1.11 | Not recorded | 6a0bee9ddbc836b87d82c492 |

### HIBCG

Final Return: 2.74. Result ID: 6a0c36bafbc373592c02db6c.

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

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

Recorded evaluation configuration: {"agents":"n=36","budget":"3M steps"}. This record does not assert independent reproduction.

### CommFormer

Final Return: 2.5. Result ID: 6a0c36bafbc373592c02db66.

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

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

Recorded evaluation configuration: {"agents":"n=36","budget":"3M steps"}. This record does not assert independent reproduction.

### QMIX

Final Return: 2.19. Result ID: 6a0c36bafbc373592c02db64.

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

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

Recorded evaluation configuration: {"agents":"n=36","budget":"3M steps"}. This record does not assert independent reproduction.

### AIB-only

Final Return: 1.87. Result ID: 6a0c36bafbc373592c02db69.

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

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

Recorded evaluation configuration: {"agents":"n=36","budget":"3M steps"}. This record does not assert independent reproduction.

### HIB-flat

Final Return: 1.72. Result ID: 6a0c36bafbc373592c02db6b.

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

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

Recorded evaluation configuration: {"agents":"n=36","budget":"3M steps"}. This record does not assert independent reproduction.

### ExpoComm

Final Return: 1.66. Result ID: 6a0c36bafbc373592c02db68.

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

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

Recorded evaluation configuration: {"agents":"n=36","budget":"3M steps"}. This record does not assert independent reproduction.

### GACG

Final Return: 1.63. Result ID: 6a0c36bafbc373592c02db67.

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

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

Recorded evaluation configuration: {"agents":"n=36","budget":"3M steps"}. This record does not assert independent reproduction.

### BVME

Final Return: 1.43. Result ID: 6a0c36bafbc373592c02db6a.

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

Recorded evaluation configuration: {"agents":"n=36","budget":"3M steps"}. This record does not assert independent reproduction.

### MAGI

Final Return: 1.11. Result ID: 6a0c36bafbc373592c02db65.

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

Recorded evaluation configuration: {"agents":"n=36","budget":"3M steps"}. This record does not assert independent reproduction.

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