It would be nice to have the “Create Image” option you can use to generate an image based on a response use a specific model instead of defaulting to Z-Image when other models are available. Could be tied to the preferred image model in image settings or let users select the model.