Today Qianwen released the Qwen-Audio-3.1 series, upgrading core ASR, TTS and Realtime interaction models and introducing new audio-creation model Qwen-Audio-3.1-TTS-Next and audio-understanding model Qwen-Audio-3.1-ASR-Next. Five new speech models form an end-to-end audio stack covering understanding, generation, interaction and creation; Qianwen cut Qwen-Audio pricing across the board—TTS down about 70%, Realtime down about 85%, ASR down about 95%—citing lower user costs.