#audio language model

StepFun unveils Step-Audio-AQAA, a fully unified audio language model that enables natural and expressive voice interaction by directly converting spoken queries into spoken responses without text intermediaries.