ResearchBenchmarksHuman Alignment Evaluation on StarCraft II (test)Human Alignment Evaluation on StarCraft II (test)Follow5Expert Bug Exploitation ScoreVLM1.1522.1513.154.149Mar 7, 2025Evaluation resultsMethodMethodLinksExpert Bug Exploitation ScoreExpert Reasoning ScoreExpert Human Simulation ScoreMid-tier Bug Exploitation ScoreMid-tier Reasoning ScoreMid-tier Human Simulation ScoreNovice Bug Exploitation ScoreNovice Reasoning ScoreNovice Human Simulation ScoreOverall Average ScoreVLM2025.0354.34.73.74.74.743.544.4MARL2025.031.321221.32.52.531.9