Flux Schnell (12B)
#24 v Text na obrázekUnknown · 2× · naposledy 30. 6. 2026
11
Momentum
FLUX.1 [schnell] is a text-to-image model developed by Black Forest Labs with 12 billion parameters, based on a "rectified flow transformer" architecture. It was trained using latent adversarial diffusion distillation and generates high-quality images in just 1 to 4 Inference steps, making it significantly faster than the Dev and Pro variants. The model is openly available under the Apache-2.0 license and can be used for personal, scientific, and commercial purposes; weights are available on Hugging Face, with reference code on GitHub.
Vývoj momenta
29.06.27.09.
Vlastnosti
| Fine-Tuning | Community fine-tunes available (70 finetune models listed on Hugging Face); official fine-tuning offered primarily for FLUX Pro |
| Generation Time | Approx. 10 seconds for a 2048px image with 4 sampling steps on a single A100 GPU |
| License | Apache 2.0 (free for commercial use) |
| Max Resolution | Tested up to approx. 2048 px (officially includes 1024x1024 and other aspect ratios like 1344x768, 896x1152) |
| Platform | Open-weight model: Hugging Face, GitHub, ComfyUI, Diffusers library, various API providers (fal, Replicate, Together AI, etc.) |
| Price | API from approx. $0.003 per megapixel (third-party, e.g. fal.ai); BFL native pricing: 1 credit = $0.01 |
| Release Date | August 1, 2024 |