ComfyUI Team just added support for Stable Diffusion 3.5 Large ControlNet models by Stability AI: Blur, Canny, and Depth—each packed with 8B parameters.

These advanced models, powered by 8 billion parameters, are available under the permissive Stability AI Community License for both commercial and non-commercial use. With these additions, users can create highly detailed and customizable images more efficiently than ever.

Overview of the New ControlNet Models

1. Blur
   • Enables ultra-high fidelity upscaling, supporting resolutions up to 8K and 16K.
   • Ideal for transforming low-resolution images into large, highly detailed visuals.
2. Canny
   • Utilizes Canny edge maps to provide structural guidance for generated images.
   • Particularly useful for creating illustrations and adaptable to various artistic styles.
3. Depth
   • Leverages depth maps from DepthFM to control the composition and perspective of generated images.
   • Perfect for applications like architectural designs, 3D asset texturing, and scenarios requiring precise spatial alignment.

Getting Started with Stable Diffusion 3.5 in ComfyUI

If you’re new to Stable Diffusion 3.5, refer to the guide for an introduction.

Try Out the ControlNet Models

To explore these new models, follow our workflows, which use the all-in-one SD3.5 large checkpoint for seamless integration. Alternatively, you can adopt the triple-clip method to load checkpoints, as detailed in earlier blog posts.

Unleash the potential of these groundbreaking models in ComfyUI and elevate your image-generation workflows!

Blur Workflow

1. Update ComfyUI to the Latest
2. Make sure the all-in-one SD3.5 large checkpoint is in your models\checkpoints folder.
3. Download sd3.5_large_controlnet_blur.safetensors and place it in your models\controlnet folder.
4. Drag and drop the image below into ComfyUI to load the example workflow

Canny Workflow

1. Update ComfyUI to the LatestMake sure the all-in-one SD3.5 large checkpoint is in your models\checkpoints folder.
2. Download sd3.5_large_controlnet_canny.safetensors and place it in your models\controlnet folder.
3. Drag and drop the image below into ComfyUI to load the example workflow, or check out Example Workflow Page.

Depth Workflow

1. Update ComfyUI to the Latest
2. Make sure the all-in-one SD3.5 large checkpoint is in your models\checkpoints folder.
3. Download sd3.5_large_controlnet_depth.safetensors and place it in your models\controlnet folder.
4. Drag and drop the image below into ComfyUI to load the example workflow (one custom node for depth map processing is included in this workflow).

Additional ControlNet models, including Stable Diffusion 3.5 Medium (2B) variants and new control types, are on the way!

Enjoy experiments with different workflows and creative possibilities!