Multi-Agent Reinforcement Learning
Benchmarks
DatasetSOTA methodMetricTrendResultsLast updated
712.4Score
96
May 4, 2026
100Win Rate (1c3s5z)
63
Jun 4, 2026
32,421Mean Episodic Reward
42
Feb 26, 2026
84Win Rate (Protoss 5 Units)
35
Jun 25, 2026
99.7Win Rate (3m)
34
Jun 18, 2026
2,757.2Average Return
24
Jun 11, 2026
3,178.7Average Return
24
Jun 11, 2026
3,676.6Average Return
24
Jun 11, 2026
-24.6Reward
24
May 14, 2026
-1.69Reward
24
May 14, 2026
36.5Reward
24
May 14, 2026
41.4Cumulative Reward
24
May 14, 2026
1.38Reward
24
May 14, 2026
0.011Collisions
23
Jun 2, 2026
0.64Collisions
23
Jun 2, 2026
0.002Collisions
23
Jun 2, 2026
6,340Mean Episodic Reward
21
Feb 26, 2026
888Mean Episodic Reward
21
Feb 26, 2026
-43Mean Episodic Reward
21
Feb 26, 2026
-161Mean Episodic Reward
21
Feb 26, 2026
-390Mean Episodic Reward
21
Feb 26, 2026
-44Mean Episodic Reward
21
Feb 26, 2026
-154Mean Episodic Reward
21
Feb 26, 2026
259.1Mean Episode Return
19
Jun 11, 2026
155.8Mean Episode Return
19
Jun 11, 2026