Official

stability-ai / stable-diffusion-3

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

Cold

Public
1.6M runs
$0.035 per image
Commercial use
GitHub
Paper
License

Run with an API

Playground API Examples README

Input

prompt

string

Shift + Return to add a new line

a photo of vibrant artistic graffiti on a wall saying "SD3 medium"a photo of vibrant artistic graffiti on a wall saying "SD3 medium"

Default: ""

aspect_ratio

string

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

Default: "1:1"

cfg

number

(minimum: 0, maximum: 20)

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

Default: 3.5

image

file

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

prompt_strength

number

(minimum: 0, maximum: 1)

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

Default: 0.85

steps

integer

(minimum: 1, maximum: 28)

Number of steps to run the sampler for.

Default: 28

output_format

string

Format of the output images

Default: "webp"

output_quality

integer

(minimum: 0, maximum: 100)

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

Default: 90

seed

integer

Set a seed for reproducibility. Random by default.

negative_prompt

string

Shift + Return to add a new line

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

Default: ""

Run this model in Node.js with one line of code:

npx create-replicate --model=stability-ai/stable-diffusion-3

or set up a project from scratch

Install Replicate’s Node.js client library:

npm install replicate

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import and set up the client:

import Replicate from "replicate";

const replicate = new Replicate({
  auth: process.env.REPLICATE_API_TOKEN,
});

Run stability-ai/stable-diffusion-3 using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const input = {
  cfg: 3.5,
  steps: 28,
  prompt: "a photo of vibrant artistic graffiti on a wall saying \"SD3 medium\"",
  aspect_ratio: "3:2",
  output_format: "webp",
  output_quality: 90,
  negative_prompt: "",
  prompt_strength: 0.85
};

const output = await replicate.run("stability-ai/stable-diffusion-3", { input });
console.log(output);

To learn more, take a look at the guide on getting started with Node.js.

Install Replicate’s Python client library:

pip install replicate

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import the client:

import replicate

Run stability-ai/stable-diffusion-3 using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "stability-ai/stable-diffusion-3",
    input={
        "cfg": 3.5,
        "steps": 28,
        "prompt": "a photo of vibrant artistic graffiti on a wall saying \"SD3 medium\"",
        "aspect_ratio": "3:2",
        "output_format": "webp",
        "output_quality": 90,
        "negative_prompt": "",
        "prompt_strength": 0.85
    }
)
print(output)

To learn more, take a look at the guide on getting started with Python.

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Run stability-ai/stable-diffusion-3 using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

curl -s -X POST \
  -H "Authorization: Bearer $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -H "Prefer: wait" \
  -d $'{
    "input": {
      "cfg": 3.5,
      "steps": 28,
      "prompt": "a photo of vibrant artistic graffiti on a wall saying \\"SD3 medium\\"",
      "aspect_ratio": "3:2",
      "output_format": "webp",
      "output_quality": 90,
      "negative_prompt": "",
      "prompt_strength": 0.85
    }
  }' \
  https://api.replicate.com/v1/models/stability-ai/stable-diffusion-3/predictions

To learn more, take a look at Replicate’s HTTP API reference docs.

Output

{
  "completed_at": "2024-06-26T18:02:19Z",
  "created_at": "2024-06-26T18:02:16.683000Z",
  "data_removed": false,
  "error": "",
  "id": "9qcv99vaxdrm00cgapvbt3fh4c",
  "input": {
    "cfg": 3.5,
    "prompt": "a photo of vibrant artistic graffiti on a wall saying \"SD3 medium\"",
    "aspect_ratio": "3:2",
    "output_format": "webp",
    "output_quality": 90,
    "negative_prompt": ""
  },
  "logs": "Seed set to: 448433150\r\nRunning workflow\r\ngot prompt\r\nExecuting node 6, title: CLIP Text Encode (Prompt), class type: CLIPTextEncode\r\nExecuting node 71, title: CLIP Text Encode (Negative Prompt), class type: CLIPTextEncode\r\nExecuting node 67, title: ConditioningZeroOut, class type: ConditioningZeroOut\r\nExecuting node 68, title: ConditioningSetTimestepRange, class type: ConditioningSetTimestepRange\r\nExecuting node 70, title: ConditioningSetTimestepRange, class type: ConditioningSetTimestepRange\r\nExecuting node 69, title: Conditioning (Combine), class type: ConditioningCombine\r\nExecuting node 135, title: EmptySD3LatentImage, class type: EmptySD3LatentImage\r\nExecuting node 271, title: KSampler, class type: KSampler\r\n  0%|          | 0/28 [00:00<?, ?it/s]\r\n  7%|▋         | 2/28 [00:00<00:01, 14.59it/s]\r\n 14%|█▍        | 4/28 [00:00<00:01, 15.14it/s]\r\n 21%|██▏       | 6/28 [00:00<00:01, 15.31it/s]\r\n 29%|██▊       | 8/28 [00:00<00:01, 15.40it/s]\r\n 36%|███▌      | 10/28 [00:00<00:01, 15.44it/s]\r\n 43%|████▎     | 12/28 [00:00<00:01, 15.47it/s]\r\n 50%|█████     | 14/28 [00:00<00:00, 15.48it/s]\r\n 57%|█████▋    | 16/28 [00:01<00:00, 15.50it/s]\r\n 64%|██████▍   | 18/28 [00:01<00:00, 15.51it/s]\r\n 71%|███████▏  | 20/28 [00:01<00:00, 15.51it/s]\r\n 79%|███████▊  | 22/28 [00:01<00:00, 15.52it/s]\r\n 86%|████████▌ | 24/28 [00:01<00:00, 15.52it/s]\r\n 93%|█████████▎| 26/28 [00:01<00:00, 15.51it/s]\r\n100%|██████████| 28/28 [00:01<00:00, 15.49it/s]\r\n100%|██████████| 28/28 [00:01<00:00, 15.44it/s]\r\nExecuting node 231, title: VAE Decode, class type: VAEDecode\r\nExecuting node 273, title: Save Image, class type: SaveImage\r\nPrompt executed in 2.28 seconds\r\noutputs:  {'273': {'images': [{'filename': 'R8_SD3_00001_.png', 'subfolder': '', 'type': 'output'}]}}\r\n====================================\r\nR8_SD3_00001_.png",
  "metrics": {
    "total_time": 2.317,
    "image_count": 1,
    "predict_time": 2.781864785
  },
  "output": [
    "https://replicate.delivery/yhqm/i4mS31BVDBYVIF0ol6UV2dwHwuztMDbsnlVJBeJ75ViVTQhJA/R8_SD3_00001_.webp"
  ],
  "started_at": "2024-06-26T18:02:16Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/9qcv99vaxdrm00cgapvbt3fh4c",
    "cancel": "https://api.replicate.com/v1/predictions/9qcv99vaxdrm00cgapvbt3fh4c/cancel"
  },
  "version": "hidden"
}

Generated in

2.8 seconds

Tweak it ShareReport

Examples

View more examples

Pricing

This model is priced by how many images are generated.

Type	Per unit	Per $1
Output	$0.035 / image or 28 images / $1

For example, generating 100 images should cost around $3.50.

Check out our docs for more information about how per-image pricing works on Replicate.

Readme

Stable Diffusion 3 Medium is a 2 billion parameter text-to-image model developed by Stability AI. It excels at photorealism, typography, and prompt following.

Stable Diffusion 3 on Replicate can be used for commercial work.

Model

Architecture diagram

Stable Diffusion 3 Medium is a Multimodal Diffusion Transformer (MMDiT) text-to-image model that features greatly improved performance in image quality, typography, complex prompt understanding, and resource-efficiency.

For more technical details, please refer to the Research paper.

Safety

As part of our safety-by-design and responsible AI deployment approach, Stability AI implement safety measures throughout the development of our models, from the time we begin pre-training a model to the ongoing development, fine-tuning, and deployment of each model. We have implemented a number of safety mitigations that are intended to reduce the risk of severe harms, however we recommend that developers conduct their own testing and apply additional mitigations based on their specific use cases.

For more about our approach to Safety, please visit our Safety page.