ResearchTasksMultistep Soft ReasoningFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedMuSRGPT-OSS-120B69Accuracy41Apr 23, 2026MUSRTMAP50.77Accuracy (Multi-choice)27May 12, 2026MuSR n=756 (test)Adaptive Local k-NN Certification (kNN no bias)100Activation (Act)11Jun 30, 2026