ResearchBenchmarksMulti-Agent Strategic Reasoning on Kuhn Poker In-domainFollow65.6Success RateGiGPO59.15260.82662.564.174Mar 4, 2026Evaluation ResultsMethodMethodLinksSuccess RateGiGPOCategory=RLCategory=RL2026.0365.6MAGECategory=Meta-RLCategory=Meta-RL2026.0365.6ReActCategory=In-Context Le...Category=In-Context Learning2026.0364.8ReflexionCategory=In-Context Le...Category=In-Context Learning2026.0364.8GRPOCategory=RLCategory=RL2026.0364.8A-MEMCategory=Memory-basedCategory=Memory-based2026.0364.1MementoCategory=Memory-basedCategory=Memory-based2026.0364.1LAMERCategory=Meta-RLCategory=Meta-RL2026.0359.4