Z-Image-Turbo
A 6B single-stream transformer distilled to 8 steps, with a Qwen3 4B text encoder and the FLUX.1 VAE. Apache 2.0, strong English and Chinese text rendering, and one of the most-downloaded model packages for ComfyUI. Tongyi Lab says it fits on 16GB consumer cards.
What you actually download
A diffusion pipeline is three separate models. These sizes are read from the real .safetensors files in the Hugging Face repository, not estimated.
Peak VRAM at 1024×1024, one image
Every combination of precision and offload strategy. Lower precision costs image quality; offloading costs speed.
| Precision | Everything in VRAM | Text encoder on CPU | Sequential CPU offload |
|---|---|---|---|
| BF16 / FP16Reference quality. What the repos ship by default. | 23.3 GB | 15.3 GB | 15.1 GB |
| FP8Roughly half the weight memory. Small quality cost, widely used for Flux and SD 3.5. | 13.1 GB | 9.1 GB | 9.0 GB |
| NF4 / INT4Aggressive 4-bit. Visible quality loss, but it puts big models on small cards. | 10.7 GB | 6.7 GB | 6.8 GB |
Which machines can run it
75 of the 79 devices we profile can run Z-Image-Turbo at 1024×1024 in some configuration. Showing the best-quality setup each one supports.
Showing the 30 devices with the most memory. The full hardware matrix covers all 79.
Licence: Apache 2.0
Yes. Z-Image-Turbo is released under Apache 2.0, which permits commercial use. Keep the licence and attribution notices with any redistribution.
Compare model licences →Z-Image-Turbo: common questions
- How much VRAM does Z-Image-Turbo need?
- At 1024x1024 and batch 1, Z-Image-Turbo peaks at about 23.3GB with everything resident at BF16, and as little as 6.7GB at NF4 + text encoder on CPU. The weights alone are 20.5GB at BF16 — 12.3GB of transformer, 8.0GB of text encoder and 0.17GB of VAE.
- Which GPUs can run Z-Image-Turbo?
- 75 of the 79 devices we profile can run it at 1024x1024 in some configuration. The Apple Mac Studio (M5 Ultra, 512GB, 2026) handles it best, at about 23.3GB. Lower precision and text-encoder offload widen that list considerably.
- Can Z-Image-Turbo be used commercially?
- Yes. Z-Image-Turbo is released under Apache 2.0, which permits commercial use. Keep the licence and attribution notices with any redistribution.
- Does offloading the text encoder help on Z-Image-Turbo?
- Yes, substantially. Z-Image-Turbo uses a large language model as its text encoder — 8.0GB at full precision. Encoding happens once per prompt, so moving it to CPU costs a couple of seconds and frees that 8.0GB for the rest of the run, taking peak VRAM at BF16 from 23.3GB down to 15.3GB.
Weight sizes read from the model repository on 2026-09-12. Activation memory is an approximation, not a benchmark — BFWR does not run a hardware test lab.
Other open image models
Guides that go deeper
Setting this model up in ComfyUI, its licence, and the hardware it needs.
- Z-Image Turbo Locally: Files, VRAM and Settings in ComfyUIThe 8-step settings, about 9GB at 8-bit, Turbo versus the base model, and Apache 2.0.Read the guide
- MOSS-Image in ComfyUI: Native Nodes, Files, Memory and SettingsThe Apache 2.0 model, the nightly-only native nodes and the noise trap.Read the guide
- How to Run Stable Diffusion Locally: Hardware, VRAM and SetupCurrent models from SD 1.5 to Qwen-Image, and the VRAM each needs.Read the guide