88Z AI, 24 September — Alibaba Tongyi Qwen has launched the Qwen-Audio-3.1 series, featuring five voice models that cover the full spectrum from understanding to creation. The series enhances multilingual and dialect recognition and transcription, improves comprehension of emotions and environmental sounds, and achieves cross-language natural timbre synthesis along with unified generation of vocals and sound effects. The real-time interaction model supports full-duplex empathetic dialogue, language switching, and Agen
At a glance
- The Qwen-Audio-3.1 series includes five models, enhancing multilingual dialect recognition, emotion understanding, and cross-language timbre synthesis.
- The Realtime model supports full-duplex empathetic interaction and agent invocation, and is compatible with hardware like smart glasses.
The update
Alibaba Tongyi Qwen has officially launched the Qwen-Audio-3.1 series, comprising five voice models that comprehensively cover scenarios of speech understanding, generation, interaction, and creation. In terms of speech recognition, the ASR model strengthens support for thirty languages and sixteen dialects, featuring role-based transcription and text polishing capabilities; the ASR-Next model further enhances generalized understanding of emotional nuances and environmental background sounds. Regarding speech generation and interaction, the TTS model supports natural timbre preservation across languages, while TTS-Next achieves unified generation of human vocals, sound effects, and ambient sounds. The Realtime model offers full-duplex interaction capabilities, enabling emotion perception, seamless language switching, and agent invocation. It has already been integrated into terminal hardware such as smart glasses, providing users with a more natural auditory experience.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- 阿里语音AISource published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.