ResearchDatasetsSayNext-BenchSayNext-BenchFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsNext-Utterance PredictionSayNext-Bench (test)29.93Top-1 Rate (GPT-4.1)6Next-Utterance PredictionSayNext-Bench 400-question (test)17.5E1 Error Rate4Next-Utterance PredictionSayNext-Bench 400-question 1.0 (test)—E1 Error0