Bring back image mixing for combining multiple images #140
I'd like to see support for 'mixing' of images — combining multiple images as inputs. Older models had this feature and we got some amazing results with it, but today's generators seem more limited this way. Wondering how many people would support this and whether it's worth pursuing.
Resolved: Several models combine up to 4 images; the picker now says so, and how to refer to each.
Good news: this is already possible, the picker just didn't say so. Several current models take more than one image and combine them: SeeDream, Nano Banana 2, ChatGPT, ChatGPT 2 and FluX 2 take up to 4, and QWEN up to 3.
With one of them selected, the image button now reads "Choose images to edit or combine" and says how many. Once two are picked, a tip shows how to refer to each in your prompt, for example: the character from image 1 in the scene of image 2.