ResearchBenchmarksMulti-Agent Reinforcement Learning on SMAC 3m to 10mvs11mMulti-Agent Reinforcement Learning on SMAC 3m to 10mvs11mFollow80Final Win RateASALT64.468.4572.576.55Jun 23, 2026Evaluation resultsMethodMethodLinksFinal Win RateASALT2026.0680Baselinetraining=from scratch...training=from scratch using MAPPO2026.0677LA-QTransformer2026.0675CORALtransfer_scheme=agent-...transfer_scheme=agent-wise transfer2026.0668CycleGANtransfer_scheme=agent-...transfer_scheme=agent-wise transfer2026.0666DANNtransfer_scheme=agent-...transfer_scheme=agent-wise transfer2026.0665