ResearchDatasetsAssistantBenchAssistantBenchFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsReward ModelingAssistantBench89.17Pairwise Accuracy13Web ReasoningAssistantBench31.8Accuracy8Agentic Task PerformanceAssistantBench (test)65.8Easy Accuracy6