SD Turbo
SD Turbo is a distilled Stable Diffusion 2.1 model released to study small, fast image generators. It can generate an image in one network evaluation, with examples for short sampling schedules.
Research release
Model design and access
The card recommends SDXL Turbo when higher image quality and prompt understanding are needed. It disables guidance and does not use negative prompts. The release is identified as a research artifact.SD Turbo official model documentation
The current Core Models page includes SD Turbo under Community and Enterprise commercial licensing. The research label describes its release purpose, not a claim that all commercial use is prohibited.Stability AI Core Models: May 2026 list
Practical use
Practical guidance. SD Turbo is useful for studying the cost of a small distilled image pipeline. Its main value is a known earlier architecture with simple sampling behavior. Compare visual quality on the actual target images before choosing it for an interactive tool; low sampling count alone does not establish the best user experience.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Access | Downloadable |
| Ancestry | Stable Diffusion 2.1 with Adversarial Diffusion Distillation.SD Turbo official model documentation |
| Preferred resolution | 512 × 512 pixels.SD Turbo official model documentation |
| Current licensing | Included in the Community and Enterprise Core Models list.Stability AI Core Models: May 2026 list |
Related models
- SDXL Turbo
Stability AI · Stable Diffusion
SDXL Turbo is an adversarially distilled SDXL 1.0 model designed for very few sampling steps.
Launch date not confirmedDownloadable