ResearchTasksOpen-endedOpen-endedFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedAlpacaEvalGPT-493.5Win Rate vs Davinci-00340Mar 24, 2026PhyX mini (test)GRPO (Fmt + Acc)2.7Overall Accuracy7Apr 16, 2026