Alibaba Qwen · First MLX build 22 Jan 2026
Qwen3-TTS-12Hz-1.7B-CustomVoice
MLX
Audio apache-2.0
A 1.9B parameter text to speech model from Alibaba Qwen, with one MLX build at 8-bit.
- Parameters
- 1.9B From Hugging Face
- Quantisation
- 8-bit Qwen3-TTS-12Hz-1.7B-CustomVoice-8bit
- Memory needed
- 5 GB Estimate, unified memory
- Download
- 2.4 GB Weights, estimated
- Downloads
- 11K This build, last 30 days
- Task
- Text to speech Audio
- MLX builds
- 1 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 5 GB at 8-bit
8 GB
16 GB
24 GB
32 GB
64 GB
128 GB
Fits with headroomTight, close other appsWon’t fit
Estimated from the size of the weights, plus a fifth for context and working memory and 2 GB for macOS and your apps. Long contexts need more.
Run it
Terminal · Hugging Face CLI$ pip install -U huggingface_hub
$ hf download mlx-community/Qwen3-TTS-12Hz-1.7B-CustomVoice-8bitMLX builds
1 on mlx-community| Qwen3-TTS-12Hz-1.7B-CustomVoice-8bit | 8-bit | 2.4 GB | 5 GB | 11K |
|---|
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.