ResearchBenchmarksMulti-Agent Reinforcement Learning on MPE Reference (test)Multi-Agent Reinforcement Learning on MPE Reference (test)Follow39.6Final Test ReturnVDN27.74430.82233.936.978Apr 19, 2026Evaluation resultsMethodMethodLinksFinal Test ReturnVDN2026.0439.6IQL2026.0437DICG2026.0435.1LLM-Priors2026.0434.4QMIX2026.0429.9DCG2026.0428.2