Model Hub

Early Alpha

The AI Image Editor is in early alpha. The model catalogue changes often.

Click the model chip in the prompt bar to open the Model Hub.

Several hundred models are available, and the header counts how many match your current filter. Search with Find a model…, or browse the categories on the left.

Categories

Group	Category	Roughly
Image generation	Image to Image	The largest group — transform an image you already have
	Text to Image	Generate from a description
	Inpainting	Fill a masked region
Video generation	Image to Video	Animate a still
	Text to Video	Generate a clip from a description
Audio & speech	Text to Audio	Music and sound effects
	Text to Speech	Voice
Tools	Image to Vector	Raster to vector
	Remove Background	Cut out the subject
	Upscale	Increase resolution

Yes — the Image Editor generates video and audio too, not only stills. The prompt bar is the same; the model decides what comes out.

Favorites collects the models you star, so the handful you actually use are one click away instead of buried in hundreds.

Reading a model card

Each card carries a sample image, the model name, a one-line description, and capability badges — TEXT TO IMAGE, PHOTOREALISM, TYPOGRAPHY, STYLIZED, TRANSFORM, REALISM and others. A green dot marks the active model. Details opens the model's own documentation.

The badges are the fastest filter in practice. If your asset is a logo with text on it, a model without TYPOGRAPHY will fight you no matter how carefully you prompt.

Choosing well

You want	Look for
Icons, sprites, UI assets	A model that can produce transparent backgrounds
Text rendered correctly in the image	TYPOGRAPHY
Photographic output	PHOTOREALISM or REALISM
A consistent house style across many assets	A style-reference model — one that accepts a saved style or reference image
Editing an existing image	The Image to Image category, not Text to Image

Set the inpainting model separately under Change inpaint model — see Editing & Layers.