Vinyals and colleagues: AlphaStar league-trained StarCraft II agent

What changed

Combined human-game imitation with a league of main and exploiter policies, reaching Grandmaster-level results in evaluated online StarCraft II games for all three races.

Scientific scope

Combined human-game imitation with a league of main and exploiter policies, reaching Grandmaster-level results in evaluated online StarCraft II games for all three races. Evaluation used stated camera and action constraints; result is not dominance in every match or transfer to arbitrary real-world multi-agent tasks.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Evaluation used stated camera and action constraints; result is not dominance in every match or transfer to arbitrary real-world multi-agent tasks.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

• Grandmaster level in StarCraft II using multi-agent reinforcement learning
• Vinyals and colleagues: AlphaStar league-trained StarCraft II agent — official research record
• Publisher-deposited complete AlphaStar byline

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction