ResearchTasksMulti-turn reasoningFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedDRIFT-BENCH (test)MUS-Repair68.71Accuracy (%)16May 26, 2026All-benchmark AverageDRIFT-5turn0.683Average Multi-turn Accuracy (multi@5)4Jun 1, 2026MT-AddRLSTA1.001Math LiC Score4Mar 6, 2026