| File | flux-2-klein-4b-Q8_0.gguf |
| Model | FLUX.2 [klein] 4B (image model, 4B) |
| Precision | Q8_0 |
| Size | 4.3 GB |
| VRAM to fit | about 6.1 GB |
| Folder | ComfyUI/models/unet/ |
| Loader node | Unet Loader (GGUF) from the ComfyUI-GGUF custom nodes |
| Source | unsloth/FLUX.2-klein-4B-GGUF on Hugging Face |
Download flux-2-klein-4b-Q8_0.gguf (4.3 GB)
Where to put it
Put the file in ComfyUI/models/unet/ (portable version: ComfyUI_windows_portable/ComfyUI/models/unet/). Press R in ComfyUI or restart it, then pick the file in Unet Loader (GGUF) from the ComfyUI-GGUF custom nodes.
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 0.1 GB: runs with part of it in system RAM, slower | flux-2-klein-4b-fp8.safetensors (4.1 GB) |
| 8 GB | Runs well | Fits in VRAM, 1.9 GB to spare | flux-2-klein-4b-fp8.safetensors (4.1 GB) |
| 10 GB | Runs well | Fits in VRAM, 3.9 GB to spare | flux-2-klein-4b.safetensors (7.8 GB) |
| 12 GB | Runs well | Fits in VRAM, 5.9 GB to spare | flux-2-klein-4b.safetensors (7.8 GB) |
| 16 GB | Runs well | Fits in VRAM, 9.9 GB to spare | flux-2-klein-4b.safetensors (7.8 GB) |
| 20 GB | Runs well | Fits in VRAM, 13.9 GB to spare | flux-2-klein-4b.safetensors (7.8 GB) |
| 24 GB | Runs well | Fits in VRAM, 17.9 GB to spare | flux-2-klein-4b.safetensors (7.8 GB) |
| 32 GB | Runs well | Fits in VRAM, 25.9 GB to spare | flux-2-klein-4b.safetensors (7.8 GB) |
| 48 GB | Runs well | Fits in VRAM, 41.9 GB to spare | flux-2-klein-4b.safetensors (7.8 GB) |
Need = file × 1 + 1 GB for the picture and 0.8 GB for Windows and the browser — the same rule as the rest of the site.
Measured with this exact file
| GPU | Time | Size | Steps | Peak VRAM | Source |
|---|---|---|---|---|---|
| RTX 5060 Ti 16 GB | 4.3 s per image | 1024x1024 | 4 | 12.4 GB | measured by me, 2026-10-01 |
Time = sampling + VAE decode, after a warm-up run. Peak VRAM = memory in use on the whole card. How I measure.
What else you need
This file alone does not make a picture. With it you load:
| Role | File | Size | Folder |
|---|---|---|---|
| Model | flux-2-klein-4b-Q8_0.gguf | 4.3 GB | models/unet |
| Text encoder | qwen_3_4b_fp8_mixed.safetensors | 5.6 GB | models/text_encoders |
| VAE | flux2-vae.safetensors | 0.3 GB | models/vae |
10.3 GB in total. Smaller text-encoder versions and other options are on the FLUX.2 klein 4B page.
Other files of FLUX.2 klein 4B
| File | Precision | Size | VRAM to fit |
|---|---|---|---|
| flux-2-klein-4b.safetensors | 16-bit | 7.8 GB | 9.6 GB |
| flux-2-klein-4b-Q8_0.gguf | Q8_0 | 4.3 GB | 6.1 GB |
| flux-2-klein-4b-fp8.safetensors | FP8 | 4.1 GB | 5.9 GB |
| flux-2-klein-4b-Q6_K.gguf | Q6_K | 3.4 GB | 5.2 GB |
| flux-2-klein-4b-Q5_K_M.gguf | Q5_K_M | 3.1 GB | 4.9 GB |
| flux-2-klein-4b-Q4_K_M.gguf | Q4_K_M | 2.6 GB | 4.4 GB |
| flux-2-klein-4b-Q3_K_M.gguf | Q3_K_M | 2.1 GB | 3.9 GB |
| flux-2-klein-4b-Q2_K.gguf | Q2_K | 1.8 GB | 3.6 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).