Qwen has announced the open-sourcing of the complete Qwen3-TTS (Text-to-Speech) model family, offering developers a versatile toolset for speech synthesis.

Model Details

The Qwen3-TTS family includes VoiceDesign, CustomVoice, and Base models. Five models are available with sizes of 0.6B and 1.8B parameters, supporting a total of ten different languages. This release includes the complete source code, pre-trained models, and several demos to facilitate usage and experimentation.

Available Resources

The models and associated resources are accessible via:
* GitHub: https://github.com/QwenLM/Qwen3-TTS
* Hugging Face: https://huggingface.co/collections/Qwen/qwen3-tts
* Blog: https://qwen.ai/blog?id=qwen3tts-0115
* Paper: https://github.com/QwenLM/Qwen3-TTS/blob/main/assets/Qwen3_TTS.pdf
* Hugging Face Demo: https://huggingface.co/spaces/Qwen/Qwen3-TTS

This release represents a significant step towards the democratization of text-to-speech technology, enabling researchers and developers to develop innovative applications in various fields.