ResearchBenchmarksMathematical Reasoning on AIME (Acc, F1, Avg. Tokens (K))Follow36.67AccuracySUPERVISORAGENT26.2728.9731.6734.37Oct 30, 2025Evaluation ResultsMethodMethodLinksAccuracyF1 ScoreAverage Tokens (K)SUPERVISORAGENTmodule=SMAS, base_agen...module=SMAS, base_agent=Smolagent2025.1036.6736.6747.95OWL2025.1033.3333.3356.11CoT SCshots=3-shotshots=3-shot2025.10303014.26Smolagent2025.10303059.14Vanilla2025.1026.6726.672.01MetaAgent2025.1026.6726.676.24