Introducing Nano Banana Pro

Google DeepMind has introduced Nano Banana Pro, a new image-generation and editing model built on Gemini 3 Pro. The model is aimed at producing accurate visuals and rendering legible text in multiple languages. Leveraging Gemini 3 Pro’s reasoning and real-world knowledge, Nano Banana Pro can pull in real-time information from Search to create context-rich outputs such as infographics, recipes and weather visualizations.

It also improves in-image text rendering and supports localization, and uses SynthID watermarks for transparency. The model handles complex compositions, blending up to 14 input images while maintaining resemblance for up to five people, and produces high-fidelity results up to 4K.

Advanced creative controls allow localized edits, camera-angle and focus changes, color grading and lighting transformations, with multiple aspect ratios and 2K/4K output options.

google deepmind, nano banana, gemini 3, image generation, image editing, text rendering, synthid, search integration, 4k output, image blending