ResearchTasksComplex Multi-step ReasoningComplex Multi-step ReasoningFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedBig-Bench HardDeepCompress-Zero-7B85.7Hard Accuracy15May 26, 2026HLEEvoMAS-760Success Rate8May 12, 2026