Google DeepMind unveils image generation tool – Nano Banana Pro

Google DeepMind launched Nano Banana Pro, the next-gen version of its earlier Nano Banana image model and the image system powering Gemini 3 Pro. The upgrade brings sharper text rendering, richer world knowledge, higher consistency across edits, and more precise creative controls — all aimed at producing studio-quality visuals from even the roughest ideas.

Why we care. Nano Banana Pro gives advertisers far more control and accuracy in generating on-brand creative, from perfectly rendered text to consistent product shots and high-quality composites — all within tools they already use like Google Ads, Slides, and Vids.

It cuts production time, improves ad relevance, and enables faster testing of creative variations, ultimately helping campaigns scale with higher-quality visuals and less manual design work.

It can:

• Generate context-rich educational visuals, grounded in real-world data via Search

• Render highly legible text in multiple languages directly inside images

• Maintain character and object consistency across up to 14 inputs

• Turn rough sketches, references, or notes into refined scenes, diagrams, mockups, and storyboards
• Perform localized edits, advanced lighting changes, and precise control over camera angles, color grading, and aspect ratios

How it works. Nano Banana Pro blends Gemini 3’s reasoning with upgraded image-editing tools to produce more accurate, on-brand visuals. Examples range from:

• Infographics and recipes grounded in real-time info
• Storyboards and architectural mockups
• Calligraphy, posters, and multilingual product packaging
• Cinematic composites merging multiple images
• High-fidelity fashion, lifestyle, and landscape renders
• Studio-level lighting transformations and focus adjustments

Who gets it. Nano Banana Pro is rolling out across Google’s ecosystem – upgraded image generation is available in Google Ads starting today.

The big picture. Nano Banana Pro moves Google’s image capabilities from quick, casual generation to professional-grade creation. With deeper reasoning, tighter control, and multilingual precision, Google is positioning the model as a creative engine for everything from classroom content to ad campaigns to full-scale film production.