Nano Banana 2.1: the image edit endpoint now takes video and PDF, not just stills.

Nano Banana 2.1: the image edit endpoint now takes video and PDF, not just stills. Google's model reads text, image, video and PDF, returns 1K/2K/4K. Same-image editing holds a character across turns, so bulk variants stop drifting. #Google #AI https://t.co/tqQjX0jCkp