Model guide · current service distinction
Stable Diffusion 3.5 Large
Stable Diffusion 3.5 Large is a model released by Stability AI. This guide separates the model's published design from what Stable Diffusion AI currently offers in its browser-based image tools.
The model behind the name
Stability AI describes SD 3.5 Large as an eight-billion-parameter model in its SD 3.5 family, designed for prompt adherence and image quality. Those are model-family descriptions, not measurements of this site's current service or guarantees for a particular image.
A model card explains supported use, setup, limitations, and version details. Hardware, output size, quality, and time to result depend on the deployment and settings. Do not infer that a web tool with 'Stable Diffusion' in its name runs this checkpoint.
Read the official model card ↗Plan the image before choosing a model
Describe the subject, visible setting, composition, light, visual medium, and any exact constraint. This portable brief is more useful than promising model-specific '8K' tags. Try it in a tool that explicitly offers SD 3.5 Large, or adapt it for our current generator with the understanding that the result may differ.
Example direction
A glass perfume bottle on pale limestone, occupying the right third; a soft side light shows the glass edge; muted sage and cream palette, editorial product photography, clear space on the left for later typography.
Inspect lettering, branding, glass edges, and proportions against any real product reference. Generated images can change facts and fine details.
Large and Turbo are different checkpoints
Stability AI describes Large as the fuller SD 3.5 model and Large Turbo as a distilled variant designed for fewer inference steps. Fewer steps can change the speed/quality tradeoff, but a model-family claim does not predict latency or quality in another service.
If choosing between them matters, use a tool that exposes both model IDs and settings, then compare your own prompts. This site currently offers neither as a selectable checkpoint.
License and output rights
The SD 3.5 model is subject to Stability AI's Community License and its conditions; do not call it unrestricted open source or assume all commercial uses are automatically cleared. This site's Terms govern use of our tools. Source images, likenesses, brands, and similar works may require separate permissions.