Google DeepMind Unveils “Nano Banana” AI Model for Advanced Image Editing

Google DeepMind has introduced a new feature that allows users to edit images using artificial intelligence, called Nano Banana. This tool forms the core of future updates for the Gemini app.

According to an official statement, Nano Banana represents the most advanced version of Gemini’s image editing capabilities. The model is designed to preserve the visual identity of a person or object with high accuracy, even when performing edits such as changing clothes, hairstyles, or other modifications.

Key Features of Nano Banana

• Enhanced realism: Nano Banana avoids common issues like “incomplete resemblance,” which can reduce the realism of personal photo edits.

• Creative freedom: Users can change outfits, locations, and even “travel through time” to recreate different eras, all while maintaining accurate facial features.

• Step-by-step editing: The model supports multi-stage editing, allowing scenes to be modified gradually. It also enables combining multiple images into a single scene and blending designs from different photos.

• Top performance: Nano Banana has achieved the highest general preference ratings compared to other models such as ChatGPT-4o, FLUX.1 Kontext, and Qwen Image Edit. It is now available to all Gemini users worldwide, both in free and paid versions.

AI Features in Google Devices

Recently, Google announced a series of advanced AI features that will first arrive on the Pixel 10 series and later roll out to older devices during the Made by Google event. These features include:

• Magic Cue: Provides contextual suggestions during conversations.

• Daily Hub: A daily summary feature.

• Journal: A diary app for daily notes.

• Text and voice-based photo editing along with camera enhancements such as Camera Coach, updates to Add Me, and Auto Best Take.

Nano Banana highlights Google DeepMind’s push toward more realistic, creative, and user-friendly AI-driven image editing.