Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

SMAC dataset and benchmarks – Page 11 · SOTA2 Research

Research Datasets

SMAC

Benchmarks

Task NameDataset NameSOTA MethodSort

Task NameDataset NameSOTA ResultTrendResults

Collaborative Multi-Agent Reinforcement Learning

SMAC Zerg (5 agents) v2

49.1Win Rate

4

Collaborative Multi-Agent Reinforcement Learning

SMAC Protoss (10 agents) v2

56.3Win Rate

4

Collaborative Multi-Agent Reinforcement Learning

SMAC Terran (10 agents) v2

58.3Win Rate

4

Multi-agent reinforcement learning on Source Tasks

SMAC Stalker-Zealot Medium-Replay quality

20.6Performance (2s3z)

4

Multi-agent reinforcement learning on Source Tasks

SMAC Stalker-Zealot Medium-Expert quality

92.5Win Rate (2s3z)

4

Multi-agent reinforcement learning on Source Tasks

SMAC Stalker-Zealot Medium quality

56.9Win Rate (2s3z)

4

Multi-agent reinforcement learning on Source Tasks

SMAC Stalker-Zealot (Expert quality)

95.6Win Rate (2s3z)

4

Offline Multi-Agent Reinforcement Learning

SMAC Marine-Hard Medium-Replay

81.9Win Rate (3m)

4

Offline Multi-Agent Reinforcement Learning

SMAC Marine-Hard Medium

84.4Win Rate (3m)

4

Multi-Agent Reinforcement Learning

SMAC Marine-Easy Unseen (test)

86.7Mean Win Rate

4

Multi-Agent Reinforcement Learning

SMAC Marine-Hard Unseen (test)

57Mean Win Rate

4

Multi-Agent Reinforcement Learning

SMAC Stalker-Zealot Seen (train)

63.4Mean Win Rate

4

Multi-Agent Reinforcement Learning

SMAC Marine-Easy Seen (train)

91.2Mean Win Rate

4

Multi-Agent Reinforcement Learning

SMAC Marine-Hard Seen (train)

79Mean Win Rate

4

Multi-Agent Reinforcement Learning

SMAC Medium-replay v2

31.3Score (Terran, 3 units)

4

Multi-Agent Reinforcement Learning

SMAC Medium v2

37.5Win Rate - Terran (3 Units)

4

Multi-agent reinforcement learning

SMAC v2 (overall)

30.3Total Mean Win Rate

4

Multi-agent reinforcement learning

SMAC v2 (seen)

31.3Terran Win Rate

4

Multi-Agent Reinforcement Learning

SMAC 8m_vs_9m v1 (test)

40.05AUC

4

Multi-Agent Reinforcement Learning

SMAC 6h vs 8z

95Win Rate

4

Multi-Agent Reinforcement Learning

SMAC 25m

100Win Rate

4

Multi-Agent Reinforcement Learning

SMAC 10m vs 11m

100Win Rate

4

Multi-Agent Reinforcement Learning

SMAC 8m vs 9m

0.95Win Rate

4

Multi-Agent Reinforcement Learning

SMAC 2s_vs_1sc (test)

1Test Win Rate

4

Sequential Multi-Agent Reinforcement Learning

SMAC 5m → 10m → 20m → 30m → 40m

83.59Win Rate

3

Page 11 of 12

Popular tasks

Multi-Agent Reinforcement Learning Multi-Agent Cooperative Control Multi-agent Micromanagement Cooperative Multi-Agent Reinforcement Learning Offline Multi-Agent Reinforcement Learning

Follow for updates

Join SOTA2 Contact