ResearchTasksComplex Instruction FollowingFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedInstructTTSEval ENGround-truth96.2APS29Jun 8, 2026InstructTTSEval-ZHGround-truth90.9APS23Jun 8, 2026in-house Chinese benchmarkSeed2.0 Pro75.26Overall Score2Jul 2, 2026Internal EvaluationDeepSeek-V4-Pro0Tie Score2Jun 19, 2026