| File | flux1-schnell-fp8-e4m3fn.safetensors |
| Model | FLUX.1 [schnell] (image model, 12B) |
| Precision | FP8 |
| Size | 11.9 GB |
| VRAM to fit | about 14.2 GB |
| Folder | ComfyUI/models/diffusion_models/ |
| Loader node | Load Diffusion Model (UNETLoader) |
| Source | Kijai/flux-fp8 on Hugging Face |
Download flux1-schnell-fp8-e4m3fn.safetensors (11.9 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 8.2 GB: runs with part of it in system RAM, slower | flux1-schnell-Q3_K_S.gguf (5.2 GB) |
| 8 GB | Offload only | Short by 6.2 GB: runs with part of it in system RAM, slower | flux1-schnell-Q3_K_S.gguf (5.2 GB) |
| 10 GB | Offload only | Short by 4.2 GB: runs with part of it in system RAM, slower | flux1-schnell-Q4_K_S.gguf (6.8 GB) |
| 12 GB | Offload only | Short by 2.2 GB: runs with part of it in system RAM, slower | flux1-schnell-Q5_K_S.gguf (8.3 GB) |
| 16 GB | Runs well | Fits in VRAM, 1.8 GB to spare | this file |
| 20 GB | Runs well | Fits in VRAM, 5.8 GB to spare | this file |
| 24 GB | Runs well | Fits in VRAM, 9.8 GB to spare | this file |
| 32 GB | Runs well | Fits in VRAM, 17.8 GB to spare | flux1-schnell.safetensors (23.8 GB) |
| 48 GB | Runs well | Fits in VRAM, 33.8 GB to spare | flux1-schnell.safetensors (23.8 GB) |
Need = file × 1 + 1.5 GB for the picture and 0.8 GB for Windows and the browser — the same rule as the rest of the site. FP8 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 | flux1-schnell-fp8-e4m3fn.safetensors | 11.9 GB | models/diffusion_models |
| Text encoder | t5xxl_fp8_e4m3fn.safetensors | 4.9 GB | models/text_encoders |
| Text encoder | clip_l.safetensors | 0.2 GB | models/text_encoders |
| VAE | ae.safetensors | 0.3 GB | models/vae |
17.4 GB in total. Smaller text-encoder versions and other options are on the FLUX.1 schnell page.
Other files of FLUX.1 schnell
| File | Precision | Size | VRAM to fit |
|---|---|---|---|
| flux1-schnell.safetensors | 16-bit | 23.8 GB | 26.1 GB |
| flux1-schnell-Q8_0.gguf | Q8_0 | 12.7 GB | 15.0 GB |
| flux1-schnell-fp8-e4m3fn.safetensors | FP8 | 11.9 GB | 14.2 GB |
| flux1-schnell-Q6_K.gguf | Q6_K | 9.8 GB | 12.1 GB |
| flux1-schnell-Q5_K_S.gguf | Q5_K_S | 8.3 GB | 10.6 GB |
| flux1-schnell-Q4_K_S.gguf | Q4_K_S | 6.8 GB | 9.1 GB |
| flux1-schnell-Q3_K_S.gguf | Q3_K_S | 5.2 GB | 7.5 GB |
| flux1-schnell-Q2_K.gguf | Q2_K | 4.0 GB | 6.3 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).