ResearchDatasetsAgentRewardBenchFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsReward ModelingAgentRewardBench83.8Precision12Web agent meta-evaluationAGENTREWARDBENCH89.21Overall Precision8Reward VerificationAgentRewardBench VisualWebArena100Precision7