ResearchTasksHACR auditHACR auditFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedAgentDojo, AgentDyn, SafeToolBench, VPIBench, VisualWebArena, DocVQA Stratified 500-taskNaive MLLM1,103Unsup. Critical Failures Count3May 20, 2026
AgentDojo, AgentDyn, SafeToolBench, VPIBench, VisualWebArena, DocVQA Stratified 500-taskNaive MLLM1,103Unsup. Critical Failures Count3May 20, 2026