Google launches Gemini 2.5 “Nano Banana” for smarter AI image editing

All edited images include a visible watermark and an invisible SynthID tag.

By Bryan Rilloraza August 30, 2025

Google has launched a new image editing model in its Gemini app called Gemini 2.5 Flash Image, nicknamed “nano banana.” Built by Google DeepMind, the model improves how users edit photos while keeping faces and features consistent.

The update focuses on likeness preservation. When users change outfits, backgrounds, or time periods, the model keeps people and pets looking like themselves. This helps avoid edits that feel off or unrealistic.

Gemini 2.5 also adds multi-step editing. Users can make changes one at a time. The model also supports style transfer, letting users apply patterns or textures from one image to another.

All edited images include a visible watermark and an invisible SynthID tag to help users and platforms identify AI-generated content.

The new features are now live in the Gemini app.

Source: 1
Featured image: Google

Bryan Rilloraza

Bryan Rilloraza has been a fixture in the local tech scene for over a decade, sharing his perspective as a tech enthusiast and industry veteran. Backed by an MBA from De La Salle University, a Bachelor’s Degree from the University of the Philippines, and 20 years of corporate experience in the telecommunications and banking sectors, Bryan provides a practical, real-world analysis of how technology serves the consumer.