Image AI
FLUX
Black Forest Labs' open and hosted model line, built by Stable Diffusion's original creators, with best-in-class photorealism
Overview
What is FLUX and what can it do?
FLUX comes from Black Forest Labs, founded by the researchers who originally built Stable Diffusion, and it has quickly become a favourite for one specific reason: photorealism that rivals or beats closed, hosted-only competitors, combined with correctly rendered hands and legible text — two areas where earlier open models consistently struggled. FLUX ships in multiple tiers: schnell, a fast open-weight model free to run and use commercially; dev, a higher-quality open-weight model restricted to non-commercial use without a separate licence; and Pro, a hosted, API-only model offering the highest overall quality for production work.
Photorealism competitive with the top closed, hosted image models
Reliable rendering of human hands and legible text, both historically weak points for open models
FLUX schnell — fast, open-weight, and free for commercial use
FLUX dev — higher-quality open weights for non-commercial or licensed use
FLUX Pro — hosted API access for the highest production-grade quality
Inpainting and structural editing tools for refining generated images
Pricing
FLUX plans and pricing in 2026
Open
$0
FLUX schnell and dev models, run locally on your own hardware
Pro API
$0.05
Hosted FLUX Pro via API, billed per generated image
Most popular
💡
Our take on pricing
schnell is genuinely free, fast, and commercially usable, making it the obvious starting point for anyone with a capable GPU. Pro API at roughly $0.05 per image is a fair price for production-grade quality without managing local infrastructure. Just remember dev requires a separate paid licence the moment you use it commercially.
Evaluation
FLUX pros and cons
Pros
- Photorealism quality among the best of any model, open or closed
- Correctly renders hands and text far more consistently than most open models
- Both a free, self-hostable option and a high-quality hosted API available
- Actively developed by the original Stable Diffusion research team
- schnell variant is fast enough for rapid iteration during creative work
Cons
- The higher-quality dev model requires a paid licence for commercial use
- Running FLUX locally at full quality needs a genuinely powerful GPU
- Smaller surrounding tooling ecosystem than the more mature Stable Diffusion
- Fewer built-in creative controls than dedicated platforms like Leonardo or Krea
Latest updates
FLUX news and recent changes
May 2026
FLUX 2 improves fine detail
The new generation sharpened output at high resolutions and improved consistency across complex scenes.
Mar 2026
Editing tools added to FLUX
Inpainting and structural composition controls were added, extending FLUX beyond pure text-to-image generation.
Verdict
Is FLUX worth it in 2026?
FLUX has earned its reputation quickly by solving two problems that plagued open image models for years: believable hands and legible text, on top of genuinely excellent photorealism. The free schnell model is a strong choice for anyone with a capable GPU wanting fast, commercially-usable results, while FLUX Pro via API delivers production-grade quality for teams that need the best output without managing their own infrastructure. The dev model's licensing requirement for commercial use is the one wrinkle to plan around, but overall FLUX is one of the strongest photorealistic options available today, open or closed.
SUBSCRIBE TO OUR PRIVATE CASES AND USEFUL TIPS
Subscribe to our newsletter, get only exclusive content and weekly digests, no any spam!
By providing my email, I accept the Privacy Policy.
FLUX Review 2026: The Complete Guide to Black Forest Labs’ Image Model
FLUX arrived with an unusual pedigree: built by the researchers who created the original Stable Diffusion after they left to start a new lab. This review examines what that experience produced, why FLUX has become a favourite specifically for photorealistic work, and how its tiered open/hosted structure works in practice.
Solving hands and text: FLUX’s signature strength
For years, garbled hands and nonsensical text were the two most reliable ways to spot an AI-generated image, regardless of which model produced it. FLUX made a deliberate engineering push to fix both, and the results are consistently more convincing than most competing models — open or closed. Combined with photorealism that holds up under close inspection, this has made FLUX a go-to choice for work where believability matters, such as product mockups, marketing visuals, and realistic character art.
Understanding the schnell / dev / Pro structure
FLUX's three-tier structure is designed to serve different needs. schnell prioritises generation speed and is fully free for commercial use, making it ideal for rapid iteration and lower-stakes work. dev produces noticeably higher quality but is licensed for non-commercial use only unless you purchase a separate commercial licence — an important distinction to plan around before using it in paid client work. Pro is hosted exclusively through the API and delivers the highest overall quality, aimed at teams that want the best output without managing GPU infrastructure themselves.
Who should use FLUX?
Developers and technical creators with a capable GPU get excellent, free, commercially-usable results from schnell with no ongoing cost.
Studios needing the highest possible quality for client-facing or production work benefit from FLUX Pro's hosted API, avoiding both infrastructure management and the dev licence question entirely.
Anyone specifically working on photorealistic content — product shots, realistic portraits, architectural visualisation — will find FLUX's strengths line up directly with their needs.
FLUX vs. Stable Diffusion and Midjourney
Stable Diffusion still has the larger, more mature tooling ecosystem and broader community support, but FLUX generally produces stronger photorealism and more reliable hands and text out of the box. Midjourney remains ahead on distinctive artistic style and ease of use for non-technical creators, but has no open-weight or self-hosting option at all. FLUX occupies a genuinely useful middle position: open enough for technical users, polished enough for production work via the Pro API.
Conclusion
FLUX in 2026 has established itself as one of the strongest photorealistic image models available, built by a team with deep first-hand experience in exactly this problem space. The free schnell model alone is a compelling reason to try it, and the option to scale up to FLUX Pro for production work without switching tools is a genuine convenience few competitors offer in the same package.
