stability-ai/stable-diffusion-3

A text-to-image model with greatly improved performance in image quality, typography, complex prompt understanding, and resource-efficiency

Capabilities

Reference ImagesNegative PromptSeed

Community model (estimated from hardware time)

Name	Type	Description	Default	Constraints
`aspect_ratio`	string	The aspect ratio of your output image. This value is ignored if you are using an input image.	`"1:1"`	1:116:921:93:22:34:55:49:169:21
`cfg`	number	The guidance scale tells the model how similar the output should be to the prompt.	`3.5`	min: 0, max: 20
`image`	string(uri)	Input image for image to image mode. The aspect ratio of your output will match this image.	`—`	—
`negative_prompt`	string	Negative prompts do not really work in SD3. Using a negative prompt will change your output in unpredictable ways.	`""`	—
`output_format`	string	Format of the output images	`"webp"`	webpjpgpng
`output_quality`	integer	Quality of the output images, from 0 to 100. 100 is best quality, 0 is lowest quality.	`90`	min: 0, max: 100
`prompt`	string	—	`""`	—
`prompt_strength`	number	Prompt strength (or denoising strength) when using image to image. 1.0 corresponds to full destruction of information in image.	`0.85`	min: 0, max: 1
`seed`	integer	Set a seed for reproducibility. Random by default.	`—`	—
`steps`	integer	Number of steps to run the sampler for.	`28`	min: 1, max: 28

aspect_ratiostring

The aspect ratio of your output image. This value is ignored if you are using an input image.

Default: "1:1"

1:116:921:93:22:34:55:49:169:21

cfgnumber

The guidance scale tells the model how similar the output should be to the prompt.

Default: 3.5min: 0, max: 20

imagestring

Input image for image to image mode. The aspect ratio of your output will match this image.

negative_promptstring

Negative prompts do not really work in SD3. Using a negative prompt will change your output in unpredictable ways.

Default: ""

output_formatstring

Format of the output images

Default: "webp"

webpjpgpng

output_qualityinteger

Quality of the output images, from 0 to 100. 100 is best quality, 0 is lowest quality.

Default: 90min: 0, max: 100

promptstring

Default: ""

prompt_strengthnumber

Prompt strength (or denoising strength) when using image to image. 1.0 corresponds to full destruction of information in image.

Default: 0.85min: 0, max: 1

seedinteger

Set a seed for reproducibility. Random by default.

stepsinteger

Number of steps to run the sampler for.

Default: 28min: 1, max: 28

Version: 527d2a6296faUpdated: 7/25/20261.9M runs