ResearchTasksAntVelocityFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedSafety GymReference policy (πref)0.93Normalized Reward6May 21, 2026Safety Gym AntVelocityReward-only5.3Computation overhead (minutes)5Jun 9, 2026