SDXL Turbo
SDXL Turbo is an adversarially distilled SDXL 1.0 model designed for very few sampling steps. Its card describes one-to-four-step generation and provides text-to-image and image-to-image examples.
Model design and access
The repository retains an older non-commercial license label and points commercial users to Stability's license page. The current Core Models list explicitly includes SDXL Turbo. Preserve the applicable license agreement when downloading or deploying the model.SDXL Turbo official model documentation
The current Core Models list includes SDXL Turbo and SD Turbo under Community and Enterprise licensing.Stability AI Core Models: May 2026 list
Practical use
Practical guidance. Turbo is a useful starting point for interactive visual studies where frequent regeneration matters. Its low-step behavior differs from normal SDXL workflows, so copying a long sampler schedule or strong guidance value can produce poor results. Review typography and faces before treating a fast preview as a finished asset.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Preferred resolution | 512 × 512 pixels.SDXL Turbo official model documentation |
| Guidance | The example disables classifier-free guidance; negative prompts are not used.SDXL Turbo official model documentation |
| Current licensing | Listed among Community and Enterprise Core Models.Stability AI Core Models: May 2026 list |
Related models
- SD Turbo
Stability AI · Stable Diffusion
SD Turbo is a distilled Stable Diffusion 2.1 model released to study small, fast image generators.
Launch date not confirmedResearch releaseDownloadable - Stable Diffusion XL 1.0
Stability AI · Stable Diffusion
SDXL 1.0 is a latent-diffusion image model that can run as a standalone base or with a separate refiner.
Launch date not confirmedAPI · Downloadable