Combining photos by hand means cutting someone out, matching scale, light and shadows, and patching the edges. In the AI Photo Editor you just upload the photos and write what you want: "put the two people from these photos side by side in an autumn park". The model works out perspective, lighting and poses so the result looks like a single real photo.
Use it to make a family photo when relatives live in different cities, a couple shot from two separate selfies, a cat on its owner's lap, or a product placed on a background from another picture. You can add up to 10 images to one edit. For people, Nano Banana 2 is the best choice — it keeps faces the most accurately; GPT Image 2.5 follows complex instructions closely, and Seedream 5 handles multiple references well.
Every result is saved as a separate image and your originals stay untouched. Refine the result with another edit, upscale it or turn it into a video. Just want a grid of photos without changing them? That's what Collage is for.