| File | z_image_int8_convrot.safetensors |
| Model | Z-Image (base) (image model, 6.15B) |
| Precision | INT8 |
| Size | 6.2 GB |
| VRAM to fit | about 8.2 GB |
| Folder | ComfyUI/models/diffusion_models/ |
| Loader node | Load Diffusion Model (UNETLoader) |
| Source | Comfy-Org/z_image on Hugging Face |
Download z_image_int8_convrot.safetensors (6.2 GB)
Where to put it
Put the file in ComfyUI/models/diffusion_models/ (portable version: ComfyUI_windows_portable/ComfyUI/models/diffusion_models/). Older guides say models/unet: that folder still works. Press R in ComfyUI or restart it, then pick the file in Load Diffusion Model (UNETLoader). Not for Mac: FP8 and INT8 files do not load on Apple Silicon.
If a workflow says “Value not in list” for that node, the workflow expects a file with another name. Click the node and choose this one.
Which graphics cards it fits
| VRAM | This file | Best file for that card (RTX 40/50) | |
|---|---|---|---|
| 6 GB | Offload only | Short by 2.2 GB: runs with part of it in system RAM, slower | z-image-Q5_K_M.gguf (5.6 GB) |
| 8 GB | Offload only | Short by 0.2 GB: runs with part of it in system RAM, slower | z-image-Q5_K_M.gguf (5.6 GB) |
| 10 GB | Runs well | Fits in VRAM, 1.8 GB to spare | this file |
| 12 GB | Runs well | Fits in VRAM, 3.8 GB to spare | this file |
| 16 GB | Runs well | Fits in VRAM, 7.8 GB to spare | z_image_bf16.safetensors (12.3 GB) |
| 20 GB | Runs well | Fits in VRAM, 11.8 GB to spare | z_image_bf16.safetensors (12.3 GB) |
| 24 GB | Runs well | Fits in VRAM, 15.8 GB to spare | z_image_bf16.safetensors (12.3 GB) |
| 32 GB | Runs well | Fits in VRAM, 23.8 GB to spare | z_image_bf16.safetensors (12.3 GB) |
| 48 GB | Runs well | Fits in VRAM, 39.8 GB to spare | z_image_bf16.safetensors (12.3 GB) |
Need = file × 1 + 1.2 GB for the picture and 0.8 GB for Windows and the browser — the same rule as the rest of the site. INT8 files do not load on a Mac (Apple Silicon); use a 16-bit or GGUF file there.
What else you need
This file alone does not make a picture. With it you load:
| Role | File | Size | Folder |
|---|---|---|---|
| Model | z_image_int8_convrot.safetensors | 6.2 GB | models/diffusion_models |
| Text encoder | qwen_3_4b_fp8_mixed.safetensors | 5.6 GB | models/text_encoders |
| VAE | ae.safetensors | 0.3 GB | models/vae |
12.2 GB in total. Smaller text-encoder versions and other options are on the Z-Image page.
Other files of Z-Image
| File | Precision | Size | VRAM to fit |
|---|---|---|---|
| z_image_bf16.safetensors | 16-bit | 12.3 GB | 14.3 GB |
| z-image-Q8_0.gguf | Q8_0 | 7.2 GB | 9.2 GB |
| z_image_int8_convrot.safetensors | INT8 | 6.2 GB | 8.2 GB |
| z-image-Q6_K.gguf | Q6_K | 6.1 GB | 8.1 GB |
| z-image-Q5_K_M.gguf | Q5_K_M | 5.6 GB | 7.6 GB |
| z-image-Q4_K_M.gguf | Q4_K_M | 5.1 GB | 7.1 GB |
| z-image-Q3_K_M.gguf | Q3_K_M | 4.6 GB | 6.6 GB |
| z-image-Q2_K.gguf | Q2_K | 4.0 GB | 6.0 GB |
Same model, different compression. 8-bit files (FP8, INT8, Q8_0) look almost the same as 16-bit; below Q5 the loss starts to show.
Want the answer for your exact card? Check your GPU. Every file of every model: ComfyUI model files.
Size and source read from Hugging Face (2026-10-01).