Z-Image
1 alias · Z-Image-Turbo, mflux 4-bit. Turbo-distilled for few-step sampling — the recommended starting point for local image generation.
Z-Image-Turbo is the fast pick of the image lane: a turbo-distilled checkpoint built for few-step sampling, so renders land in seconds. It serves through the OpenAI-compatible /v1/images/generations endpoint — the same server the desktop app's Images tab drives — as a 5.9 GB mflux 4-bit conversion that runs on a 16 GB Mac.
- family
- Z-Image
- aliases
- 1
- install
- pip install 'rapid-mlx[image]'
- OpenAI base URL
- http://localhost:8000/v1
Usage
pip install 'rapid-mlx[image]'
rapid-mlx serve z-image-turbo
curl http://localhost:8000/v1/images/generations \
-H 'Content-Type: application/json' \
-d '{"model":"z-image-turbo","prompt":"a lighthouse at dusk, oil painting","size":"1024x1024"}'
Download
Mirrored on the rapid-mlx CDN — our fresh-install test measured pulls at 90 MB/s, with automatic mid-pull fallback to Hugging Face if a mirror file slows down. One command, no account:
rapid-mlx pull z-image-turboZ-Image-Turbo · 5.9 GB · mirrored ✓ · live status
Weights land in the standard Hugging Face cache, and rapid-mlx serve pulls automatically on first use.
Aliases
| alias | hf repo | min RAM | notes |
|---|---|---|---|
| z-image-turbo | filipstrand/Z-Image-Turbo-mflux-4bit | 16 GB | Recommended starting point. Turbo-distilled — renders in seconds. |
Frequently asked questions
Can I run Z-Image locally on a Mac?
Yes — z-image-turbo renders entirely on Apple Silicon through POST /v1/images/generations. 5.9 GB download, 16 GB Mac, no cloud.
How fast is Z-Image-Turbo?
Turbo-distilled for few-step sampling — seconds per render rather than tens of seconds.