- Home
- Text-to-speech tools
- Qwen3-TTS
Qwen3-TTS
What it is
Qwen3-TTS is an open-source series of text-to-speech models from Alibaba Cloud’s Qwen team. It generates speech in streaming or non-streaming mode and supports custom voices, voice design from natural-language descriptions, voice cloning from reference audio, and natural-language voice control. The released models include 0.6B and 1.7B parameter variants for custom voice or base tasks, plus a 1.7B voice design model. The listed language count is 10, and WAV is an export format. The project provides a Python package and model downloads through Hugging Face and ModelScope, as well as a local Gradio interface. It recommends a fresh Python 3.12 environment; its examples use CUDA, and FlashAttention 2 is recommended to reduce GPU memory use on compatible hardware. The README reports synthesis latency as low as 97 ms. Alibaba Cloud DashScope APIs are linked for hosted voice features and require an API key. The project is licensed under Apache-2.0 and permits commercial use.
For and against
For
- Open-source under Apache-2.0; commercial use is permitted
- Supports streaming and non-streaming generation
- Offers voice design and reference-audio cloning
- Lists 10 languages and WAV export
- Includes a local Gradio web demo
Against
- Local setup recommends a fresh Python 3.12 environment
- CUDA examples and FlashAttention 2 depend on compatible hardware
- Cloning quality may fall without a reference transcript
- vLLM-Omni online serving is planned, not currently supported
The verdict
Qwen3-TTS combines speech generation with voice design and cloning in an open-source model series. Check the Python and hardware requirements, or the API-key requirement for hosted DashScope use, before choosing a deployment route.
Who it suits
It suits developers and teams looking for open-source speech generation, voice design, or cloning tools for local or API-based use. The project’s setup guidance is oriented toward Python environments and compatible GPU hardware.
Plans and prices
Qwen3-TTS open-source models
0.6B and 1.7B models · 10 languages
Read 7 Oct 2026 · github.com
On the record
- API access
- Yesgithub.com · read 20 Sep 2026
- API integration
- The project links Alibaba Cloud DashScope real-time APIs for custom voice, voice cloning, and voice design.github.com · read 7 Oct 2026
- API pricing
- Alibaba Cloud lists international Qwen3-TTS API rates of $0.115 per 10,000 input characters for Voice Design and Voice Cloning, with output free; the page also lists a 110,000-character free quota for each.alibabacloud.com · read 7 Oct 2026
- API requirements
- The linked Alibaba Cloud real-time synthesis guide says API use requires configuring an API key and installing the latest DashScope SDK when using that SDK.help.aliyun.com · read 7 Oct 2026
- Commercial use
- Yesgithub.com · read 20 Sep 2026
- Download and install
- Models can be downloaded through Hugging Face or ModelScope, and the project provides a `qwen-tts` Python package installable from PyPI.github.com · read 7 Oct 2026
- Export formats
- WAVgithub.com · read 20 Sep 2026
- Generation modes
- The models support streaming and non-streaming speech generation, voice design, voice cloning, and natural-language voice control.github.com · read 7 Oct 2026
- Hardware and setup
- The README demonstrates model loading on CUDA and recommends FlashAttention 2 to reduce GPU memory use, subject to compatible hardware.github.com · read 7 Oct 2026
- Inference integration
- The README states that vLLM-Omni supports offline Qwen3-TTS inference and that online serving is planned for later support.github.com · read 7 Oct 2026
- Languages
- 10github.com · read 20 Sep 2026
- Languages
- The project lists support for Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, and Italian, as well as multiple dialectal voice profiles.github.com · read 7 Oct 2026
- Latency
- The README says the streaming architecture can produce the first audio packet after one character and reports end-to-end synthesis latency as low as 97 ms.github.com · read 7 Oct 2026
- License
- The GitHub repository lists the Apache-2.0 license.github.com · read 7 Oct 2026
- Local installation
- The project provides a Python package install and says it recommends a fresh Python 3.12 environment.github.com · read 7 Oct 2026
- Local web interface
- A local Gradio web UI demo can be launched with the qwen-tts-demo command.github.com · read 7 Oct 2026
- Model hosting
- The README links model downloads on Hugging Face and ModelScope and documents loading weights locally or by model ID.github.com · read 7 Oct 2026
- Model sizes
- Released models include 0.6B and 1.7B parameter variants for custom voice or base tasks, plus a 1.7B voice design model.github.com · read 7 Oct 2026
