Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-agent reinforcement learning on SMAC v2 (5v5 and 5v6 Matchups)
Loading...
89
Protoss 5v5 Win Rate
MARL-GPT
55.72
64.36
73
81.64
Apr 7, 2026
Protoss 5v5 Win Rate
Protoss 5v6 Win Rate
Terran 5v5 Win Rate
Terran 5v6 Win Rate
Zerg 5v5 Win Rate
Zerg 5v6 Win Rate
Updated 11d ago
Evaluation Results
Method
Method
Links
Protoss 5v5 Win Rate
Protoss 5v6 Win Rate
Terran 5v5 Win Rate
Terran 5v6 Win Rate
Zerg 5v5 Win Rate
Zerg 5v6 Win Rate
MARL-GPT
Training Scope=Multi-E...
2026.04
89
54
93
63
74
46
Expert
Task Specificity=Singl...
2026.04
87
49
91
61
72
48
BC-LSTM
Training Scope=Single-...
2026.04
85
30
88
51
72
38
DT
Training Scope=Single-...
2026.04
82
30
84
48
65
34
RATE
Training Scope=Single-...
2026.04
79
28
85
41
64
33
BC
Training Scope=Single-...
2026.04
61
12
69
24
56
24
CQL
Training Scope=Single-...
2026.04
57
14
69
28
50
23
Feedback
Search any
task
Search any
task
