ResearchBenchmarksCode Generation on MBPP (Acc, F1, Avg. Tokens (K))Follow85.68AccuracySmolagent78.81680.59882.3884.162Oct 30, 2025Evaluation ResultsMethodMethodLinksAccuracyF1 ScoreAverage Tokens (K)Smolagent2025.1085.6885.68111.07SUPERVISORAGENTmodule=SMAS, base_agen...module=SMAS, base_agent=Smolagent2025.1084.4384.43103.71CoT SCshots=3-shotshots=3-shot2025.1081.2681.261.29Vanilla2025.1080.0980.090.27MetaAgent2025.1079.8679.866.39OWL2025.1079.0879.0854.8