FLUX.1 dev needs about 9.8 GB of memory at Q5_0 (a 8.3 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 11 quants, from 4.0 GB (Q2_K) to 23.8 GB (F16).
FLUX.1 dev – 11.9B parameter diffusion model in GGUF format
FLUX.1 dev is a 11.9 billion‑parameter diffusion model converted to GGUF. It supports image generation via ComfyUI‑GGUF custom node. The model is available in multiple quantizations, from Q2_K to F16.
Good at
- Image generation
- Low‑latency inference on modest GPUs
- Versatile quantization options
Limits
- No vision support in text‑only contexts
- License terms inherited from parent model
- Requires 8 k context for memory estimates
Best for
- Generating images from prompts
- Fine‑tuning on custom datasets
- Running on consumer GPUs with 8 GB VRAM
Who it’s for. Developers and researchers who need a diffusion model in GGUF format for integration with ComfyUI or llama.cpp. The model is suitable for those who prefer quantized weights to reduce VRAM usage.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 4.0 GB | 5.0 GB | Fits | Get file → | |
| Q3_K_S | 5.2 GB | 6.4 GB | Fits | Get file → | |
| Q4_0 | 6.8 GB | 8.1 GB | Too big | Get file → | |
| Q4_K_S | 6.8 GB | 8.1 GB | Too big | Get file → | |
| Q4_1 | 7.5 GB | 8.9 GB | Too big | Get file → | |
| Q5_0 | 8.3 GB | 9.8 GB | Too big | Get file → | |
| Q5_K_S | 8.3 GB | 9.8 GB | Too big | Get file → | |
| Q5_1 | 9.0 GB | 10.6 GB | Too big | Get file → | |
| Q6_K | 9.9 GB | 11.5 GB | Too big | Get file → | |
| Q8_0 | 12.7 GB | 14.7 GB | Too big | Get file → | |
| F16 | 23.8 GB | 27.2 GB | Too big | Get file → |
For 8 GB VRAM, Q4_0 or Q4_K_S (~8.1 GB memory) is recommended. If 12 GB VRAM is available, Q5_0 or Q5_K_S (~9.8 GB) offers higher quality. For maximum fidelity, use F16 (~27.2 GB) on high‑end GPUs.
What GPU do you need for FLUX.1 dev?
These figures add the file size, an fp16 KV cache at an 8K context window and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: no quant fits at an 8K context.
- 6 GB VRAM: the best quant that fits is Q2_K (4.0 GB file, about 5.0 GB in use).
- 8 GB VRAM: the best quant that fits is Q3_K_S (5.2 GB file, about 6.4 GB in use).
- 12 GB VRAM: the best quant that fits is Q6_K (9.9 GB file, about 11.5 GB in use).
- 16 GB VRAM: the best quant that fits is Q8_0 (12.7 GB file, about 14.7 GB in use).
- 24 GB VRAM: the best quant that fits is Q8_0 (12.7 GB file, about 14.7 GB in use).
- 32 GB VRAM: the best quant that fits is F16 (23.8 GB file, about 27.2 GB in use).
Which FLUX.1 dev quant should you choose?
Q5_0 (8.3 GB) is the middle option in this repository. Closer to the original model than 4-bit, for a moderate increase in size. A good pick when you have memory to spare.
If memory is tight, the smallest file is Q2_K at 4.0 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about FLUX.1 dev
How much VRAM does FLUX.1 dev need?
About 9.8 GB at Q5_0 with an 8K context window. That is the 8.3 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run FLUX.1 dev on an 8 GB GPU?
Yes, with Q3_K_S: the 5.2 GB file needs about 6.4 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of FLUX.1 dev GGUF?
It depends on the quant: Q2_K is 4.0 GB, Q3_K_S is 5.2 GB, Q4_0 is 6.8 GB, Q4_K_S is 6.8 GB, Q4_1 is 7.5 GB, Q5_0 is 8.3 GB and 5 more.
Where can I download FLUX.1 dev GGUF?
The files are on Hugging Face in the city96/FLUX.1-dev-gguf repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does FLUX.1 dev use?
Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run FLUX.1 dev on 8 GB VRAM?
Yes, using Q4_0 or Q4_K_S quantizations, which estimate 8.1 GB memory at 8 k context.
What is the file size of the smallest quantized model?
The Q2_K quantization is 4.03 GB.
Where do I place the model files for ComfyUI?
Place them in ComfyUI/models/unet as instructed in the repo readme.