FLUX.1 [schnell]
Black Forest Labs · Image generation · 12B · Released 1 August 2024
FLUX.1 schnell is the fast, permissively-licensed member of the FLUX family. It is distilled to generate high-quality images in as few as one to four steps, which keeps it quick even on mid-range hardware, and unlike its dev sibling it is Apache 2.0, so there are no commercial-use conditions. The trade-off is a little quality and prompt fidelity versus dev, but for most uses, and especially for commercial or high-volume work, schnell is the practical FLUX to run. It is well supported in ComfyUI and InvokeAI.
Strengths
- Very fast generation, one to four steps
- Renders readable text in images, a weak point for older models
- Apache 2.0, so fully usable commercially
Weaknesses
- Slightly lower peak quality than FLUX.1 [dev]
- A 12B model needs 12GB or more, or quantisation, to run comfortably
- A smaller fine-tune and LoRA ecosystem than SDXL
Hardware requirements
| Quantisation | Approx. VRAM | Notes |
|---|---|---|
| GGUF Q4 | ~8GB | Quantised with offloading, fits an 8GB to 12GB card |
| FP8 | ~12GB | A good balance on a 16GB card |
| FP16 | ~24GB | Full precision, including the text encoder |
Optimised builds available for Apple Silicon.
What you'd need to run this
Roughly what a machine to run this would need, at up to three levels of quality. Memory is the deciding factor.
Minimum to run it
GGUF Q4 · ~8GB needed
One 12GB GPU
NVIDIA GeForce RTX 3060 12GBor a Mac or mini-PC with unified memory, if you prefer no discrete GPU, Mac mini M5 Pro .
At least 32GB of system RAM alongside the card.
around £700–£1,100
What else 12GB runs →For good quality
FP8 · ~12GB needed
One 16GB GPU
Intel Arc A770 16GBor a Mac or mini-PC with unified memory, if you prefer no discrete GPU, Mac mini M5 Pro .
At least 32GB of system RAM alongside the card.
around £700–£1,100
What else 16GB runs →Best quality
FP16 · ~24GB needed
One 32GB GPU
NVIDIA GeForce RTX 5090or a Mac or mini-PC with unified memory, if you prefer no discrete GPU, Mac mini M4 Pro .
64GB of system RAM alongside the card.
Licence
Apache 2.0 — read the licence
FLUX.1 [schnell]: common questions
- What hardware do I need to run FLUX.1 [schnell]?
- At its most compressed (GGUF Q4) it needs roughly 8GB of VRAM, and about 16GB for good quality. VRAM figures are approximate and depend on context length and settings.
- Is FLUX.1 [schnell] free for commercial use?
- Yes. FLUX.1 [schnell] is licensed under Apache 2.0, which permits commercial use with no meaningful conditions.
- Can I run FLUX.1 [schnell] on Apple Silicon?
- Yes. FLUX.1 [schnell] has builds optimised for Apple Silicon, through MLX or GGUF on a Mac.
Availability
Recommended for
- Fast local image generation on a mid-range card
- Commercial work that needs a permissive image model
- Images with text, such as posters or UI mockups
Related models
Related guides
Glossary
Catalogue entry last verified 30 July 2026. Specifications change; verify anything you are about to spend money on.