FLUX.1 [Schnell]
Black Forest Labs' fastest open-weights image model. Apache-2.0 licensed, ~1-4 step inference.
FLUX.1 [Schnell] is image generation AI model from Black Forest Labs, priced at €0.000 per 1M input tokens with a unknown context window.
Pricing
API Integration
Use our OpenAI-compatible API to integrate FLUX.1 [Schnell] into your application.
npm install railwailimport railwail from "railwail";
const rw = railwail("YOUR_API_KEY");
const images = await rw.run("flux-1-schnell-apache", "A beautiful sunset over Tokyo");
console.log(images[0].url);
// Or use the image() method for full control
const res = await rw.image("flux-1-schnell-apache", "A cat in space", {
size: "1024x1024",
n: 1,
});
console.log(res.data[0].url);Deep dive — Black Forest Labs's FLUX.1 [Schnell]
Black Forest Labs was founded in August 2024 by Robin Rombach, Patrick Esser, Andreas Blattmann and Dominik Lorenz — the team that created Latent Diffusion (CVPR 2022) at LMU Munich and led Stable Diffusion development at Stability AI. The lab is headquartered in Freiburg, Germany and raised a $31M seed round led by Andreessen Horowitz in 2024. Their FLUX.1 family is the European answer to DALL-E 3, Imagen and Midjourney, and FLUX.1 [schnell] is the only open-weights member released under the permissive Apache 2.0 license, making it the de-facto baseline for the open-source image-generation community.
Visit Black Forest Labs →FLUX.1 [schnell] is the distilled, fast inference variant of the FLUX.1 family released by Black Forest Labs in August 2024. The underlying architecture is a rectified-flow Diffusion Transformer (DiT) with approximately 12B parameters, trained with the flow-matching objective described in the SD3 paper (Esser et al. 2024). It uses T5-XXL plus CLIP-L as text encoders and a 16x learned latent VAE for image tokens. The 'schnell' variant is produced by adversarial latent-space distillation (related to the SDXL Turbo / Adversarial Diffusion Distillation recipe by Sauer et al. 2023) so that 1-4 sampling steps suffice instead of the 28-50 of the dev/pro versions. Crucially, FLUX.1 [schnell] is the only FLUX.1 model released under the permissive Apache 2.0 license, which allows unlimited commercial use, redistribution and derivative works. This has made it the dominant open-weights image model since late 2024.
- Parameters
- ~12B parameters
- Context
- 512 tokens
- Open-weights under Apache 2.0 — fully commercial use permitted
- 1-4 step sampling for very fast inference
- Photorealism and prompt adherence competitive with closed peers
- Up to 2 MP output resolution
- Excellent typography for an open model (T5-XXL conditioning)
- Self-hostable on a single H100/A100
- Compatible with ComfyUI, Diffusers, LoRA fine-tuning, ControlNets
- Best for: self-hosted creative apps, batch generation, on-prem deployments, fine-tuning.
Pretrained on a large licensed and publicly available image-text corpus, then distilled to few-step inference using adversarial latent-space techniques. Exact data and step counts are not disclosed.
License: Apache 2.0 — fully open, commercial use permitted including redistribution and derivatives.
Known limitations
- Lower quality than FLUX.1 [pro]/[dev] in fine details
- Few-step sampling can hurt complex compositions
- No built-in safety classifier in the weights
- Hands and small text still imperfect
Frequently asked questions
Related Models
View all Image GenerationFlux 1.1 Pro Ultra
FLUX 1.1 Pro in ultra mode. Up to 4 megapixel images with raw mode for photorealism.
Flux Dev
Black Forest Labs' development model. Fast, high-quality image generation with LoRA support.
Google Imagen 4
Google's Imagen 4. Text-to-image with strong photorealism and improved typography support.
Google Imagen 4 Ultra
Premium Imagen 4 tier. Highest fidelity, prompt adherence and typography quality from Google.
Start using FLUX.1 [Schnell] today
Get started with free credits. No credit card required. Access FLUX.1 [Schnell] and 100+ other models through a single API.