ResearchTasksSecurity AnalysisFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedSecurity Tasks 15,000 benign samples (test)SecureCAI96.8F1 (Log)5Feb 26, 2026CyberGymSageAgent60.2Resolved Percentage4Feb 26, 2026ConTesterCellSpecSec-ARI100Accuracy2Feb 26, 2026