SDXL Turbo Online

SDXL Turbo is the breakthrough text-to-image model that does in one network evaluation what used to take 50 steps. Using Adversarial Diffusion Distillation (ADD) for training, it's a faster, distilled version of SDXL 1.0 that synthesizes high-quality, photorealistic images from text prompts in real time — without sacrificing the visual quality that made SDXL impressive in the first place.

For developers, researchers, and creators who need rapid visual iteration, SDXL Turbo removes the wait that used to be part of the process. Real-time image synthesis, unprecedented quality, one step. The future of fast generative AI is already here.