Best-in-class text rendering
Renders legible, correctly-spelled text inside images — historically the hardest thing for image models.
Top open-weights image generation model with best-in-class text fidelity.
FLUX 1.1 Pro is Black Forest Labs' top open-weights image model, released in August 2025. It leads on text rendering and hand fidelity — the two areas where earlier image models most visibly failed.
Because the weights are open, it slots into self-hosted, production inference pipelines and has replaced SDXL in most production runs. It generates images only — for video or audio you pair it with a model like Sora 2.
Renders legible, correctly-spelled text inside images — historically the hardest thing for image models.
Self-hostable weights for production pipelines, with no per-image API lock-in.
Stable, fast inference that drops into existing generation pipelines at scale.
Closely follows complex prompts for composition, style, and subject detail.
Generating on-brand product and lifestyle shots at scale without a photo shoot.
2Kresolution outputIterating on ad and social visuals — including images with legible headline text.
textrendered legibly in-imageEmbedding self-hosted image generation directly into a product feature.
openweights for self-hostingNo video or audio output. For motion or sound you need a separate model such as Sora 2.
Quality scales with prompt craft. Getting consistent, on-brand output takes prompt engineering and often a reference workflow.
You run the GPUs. Open weights mean you also own the inference infrastructure and its cost.
Yes — its weights are open, so it can be self-hosted in production pipelines rather than only called via an API.
Yes — legible in-image text is one of its headline strengths, where earlier models like SDXL struggled.
No — it generates images only. Pair it with a video model such as Sora 2 for motion.
Our weekly AI brief — written by the team shipping it.
Joined by 4,200+ engineers, founders & product leads · Unsubscribe anytime