# Alibaba Tongyi Qianwen Launches Speech Synthesis Large Model Qwen-Audio-3.0-TTS

- Category: Flash (快讯)
- Published: 2026-07-21T11:52:00.000Z
- Source date: 2026-07-21
- HTML: https://www.88z.ai/en/news/flash-20260721-e34de92ac05b720d
- Markdown: https://www.88z.ai/en/news/flash-20260721-e34de92ac05b720d.md
- Other language: https://www.88z.ai/news/flash-20260721-e34de92ac05b720d.md

Alibaba has officially launched Qwen-Audio-3.0-TTS, a large speech synthesis model featuring Flash for real-time interaction and Plus for high-quality generation. The model delivers systematic improvements in fine-grained tag control, natural language style definition, multilingual and dialect coverage, and acoustic robustness. It supports precise tone and emotion regulation via structured tags in text, covering 16 languages and 20 dialects.

## At a glance

- Launches dual versions: Flash for real-time interaction and Plus for high-quality generation
- Supports 16 languages, 20 dialects, and structured emotion control

## The update

Alibaba has officially released the new large speech synthesis model, Qwen-Audio-3.0-TTS. The model offers two primary versions: the Flash version optimized for real-time interaction scenarios, and the Plus version focused on high-quality audio generation, catering to diverse user needs. In terms of technical performance, the new model achieves systematic upgrades. It enhances fine-grained tag control capabilities, allowing users to define styles via natural language instructions, while improving robustness in complex acoustic environments. Furthermore, the model supports embedding structured tags within text to precisely regulate tone and emotion. Its language coverage is extensive, supporting 16 languages and 20 dialects, significantly boosting multilingual synthesis expressiveness.

## Sources

- [千问大模型](https://mp.weixin.qq.com/s/ZhDZfqf6IFKnv8qU2LW1DA) (2026-07-21)

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.