ResearchTasksSafe Tool ExecutionSafe Tool ExecutionFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedAethelgard N=500 BaselineBaseline (raw)5.3Safe Execution Rate (SER)1Apr 15, 2026Aethelgard Real SplitAethelgard337SER Improvement (Real)1Apr 15, 2026Aethelgard Synthetic SplitAethelgard260SER Improvement (Synthetic)1Apr 15, 2026