We have been hearing about Nano Banana for the past few days. It was an incredible model that could produce new visuals from a given image while maintaining amazing character consistency. As it turns out, this was Gemini 2.5 Flash all along. It is without a doubt one of the most impressive image models we have tested. You can use a single image to change angle, environment, and other elements in our image.
🖼️ Creative composition
Combine creative elements from multiple images with a single prompt. With 2.5 Flash, you can start blending different elements from up to three inputs to create a unique, unified composition pic.twitter.com/WVzlkz0Cbp
— Google DeepMind (@GoogleDeepMind) August 26, 2025
2.5 Flash can transform styles from one image to another while preserving subject’s form and details. It is also possible to combine images from. You can even ask Gemini to generate what happens after a given scenario in an image. Not surprisingly, Gemini 2.5 Flash Image now leads image editing by a big distance over GPT-4o.
Google’s Gemini 2.5 Flash Image (Nano-Banana) takes the crown as the leading image editing model, beating GPT-4o and Qwen-Image-Edit in the Artificial Analysis Image Editing Arena!
We were given early access and have been testing it in our arena under the pseudonym ‘rex’ for the… pic.twitter.com/Aa15nLgjZt
— Artificial Analysis (@ArtificialAnlys) August 26, 2025
[HT]
