Google improves Gemini AI image editing with “nano banana” model

In a remarkable development in AI-driven image editing, Google's latest model, dubbed 'nano banana,' has taken the spotlight. This innovative technology has swiftly ascended to the top of the LMArena leaderboard, showcasing its impressive capabilities. Google has announced that this cutting-edge model, developed by Google DeepMind, is now available in the Gemini app. The introduction of AI image editing allows users to modify pictures simply by entering prompts, eliminating the need for complex software like Photoshop. Earlier this year, Google debuted editing features in Gemini, which proved to be quite effective. However, as with many generative systems, the unpredictable nature of these tools often led to unexpected changes in image elements. The nano banana model, technically known as Gemini 2.5 Flash Image, boasts unparalleled consistency in edits. Unlike its predecessors, it has the ability to recall details, ensuring that users achieve more reliable results with each modification. This opens up exciting possibilities for AI image editing. Users can upload photos and transform a person's appearance, experimenting with different styles or outfits. For instance, you might envision someone as a matador or a character straight out of a '90s sitcom. Thanks to the consistent nature of the nano banana model, the resulting images will still bear a strong resemblance to the original subjects, even after multiple edits. Google assures users that the integrity of the source material remains intact throughout the editing process.