Input batches

It's free to upload and store input images, which will be used to train your Stable Diffusion model.

  • up to 50 images per input batch


Training runs of Stable Diffusion Dreambooth cost $5.00 each.

  • up to 5000 steps per finetune

We'll store your finetune for whenever you want to create generations (output batches) from it.

You can delete your finetune at any time. Finetunes that haven't been used in 30 days will be deleted automatically.

Output batches

Output generations on your finetuned model cost $0.50 each.

  • up to 8 images per prompt
  • up to 8 prompts
  • up to 75 denoising steps per image

If you submit a request for 8 prompts, with 8 images each, then your output batch will contain 64 images total, and it will cost $0.50.

You can delete your output batch at any time.

Generated images will be stored on our CDN, which you can use directly in your app for fast image loading.