ResearchBenchmarksMonkey-banana task execution on Dual Bananas Scene 5Follow95Success RatePlanORN48.260.3572.584.65Feb 9, 2026Evaluation ResultsMethodMethodLinksSuccess RateAverage StepsAverage Time (s)Average TokensTime to SuccessTokens to SuccessPlanORN2026.029542.5203.04—213.72—OR2026.029422.462.41—66.39—ORN2026.02932384.25—90.59—ReAct2026.028828.6247.76—281.55—MLDT2026.026517.339.31—60.48—Basic2026.025019.634.89—69.77—