ResearchBenchmarksAgentic Task Solving on AppWorld (test-n)Follow81.15TGC AverageCLEAR71.654874.119976.58579.0501Apr 8, 2026Evaluation ResultsMethodMethodLinksTGC AverageTGC Pass@3SGC AverageSGC Pass@3CLEARModel for πE=Claude-So...Model for πE=Claude-Sonnet-42026.0481.1591.6766.6782.14ACEModel for πE=Claude-So...Model for πE=Claude-Sonnet-42026.0474.485.7158.9373.21BaselineModel for πE=Claude-So...Model for πE=Claude-Sonnet-42026.0472.6286.952.3866.07RAGModel for πE=Claude-So...Model for πE=Claude-Sonnet-42026.0472.0286.3154.6771.43