Image Generation

Image generation can be accomplished with compatible models.

Set 'Image' Purpose

Most models capable of image generation will come with image already enabled as a model's purpose.

However, if you are using a model, or have added a custom model, where you know that the model supports image generation, then you'll need to enable image as a purpose for the model.

Adding custom model

When adding a custom model, select image as a purpose.

Updating existing model

To update purposes for an existing model, in the conversation area, click on the model selector, find the model you want to update, and then click on the edit icon.

This will open a modal where you can update the model's purpose.

Be sure to appropriately select purposes based on if the model is image-only or if it is multimodal.

Image-only models

It is very important that if the model is an image-model only, meaning the model doesn't have multimodal functionality, then ensure that only image is selected.

Multimodal models

Likewise, if the model has multimodal functionality, then ensure that more than just image is selected as a purpose. Such as, most multimodal models that support images are also likely to support text.

Advanced configurations

Not all models support advanced configurations - such as, these configurations may likely work for OpenAI models but may not work for Gemini.

Advanced configurations can be included as model parameters.

Aspect Ratio

size : 1536x1024

Number of images generated

n: 2