Google DeepMind has introduced a new Nano Banana image editing model inside the Gemini app, aiming to give users better control over their photos while keeping likeness consistent. The update follows earlier launches of native editing tools in the app and builds on feedback that even small changes to familiar faces can feel off.
Why it Matters: As more people turn to AI for creative edits, keeping personal likeness accurate is important for trust and usability. By focusing on preserving how people and pets actually look, Google hopes the Gemini app can stand out as a reliable tool for everyday photo editing.
The new Nano Banana model allows users to upload photos and request changes ranging from trying new outfits and hairstyles to previewing room designs. The system is designed to maintain consistency across edits, whether placing yourself in a new location, blending multiple photos together, or adjusting only certain elements of an image.
Among the key features are:
Costume or location changes while keeping faces and appearances intact.The ability to blend multiple photos into a single scene.Multi-turn editing that lets users adjust parts of an image step by step.Design mixing, such as applying textures or patterns from one image onto objects in another.
All edited images in Gemini will carry a visible watermark, along with Google’s invisible SynthID watermark, to indicate they are AI-generated. The updated image editing capabilities are available starting today in the Gemini app.
Do you see yourself using Gemini for creative edits or everyday photo adjustments?