Google unveils Nano Banana Pro image model

Google has launched Nano Banana Pro, a next‑generation image model built on Gemini 3 that targets professional image creation and editing. The update emphasizes higher fidelity outputs (including 4K), stronger text rendering accuracy, and “world knowledge” via integration with Google Search so the model can pull web data for accurate labels, infographics, and complex graphic layouts.

Key capabilities called out by Google include the ability to use up to 14 visual references in a single composition, preserve character identities across up to five people, and finer control over photographic attributes like camera angle, focus, and lighting. Nano Banana Pro also improves on text handling: it supports long text inputs, multiple languages, fonts, and structured graphic layouts—addressing a major pain point for using AI images in real work.

Google and The Rundown suggest practical use cases such as precise infographics, storyboards, and high‑converting lead magnets. To try it, the tutorial recommends opening the Gemini app (mobile or web), going to chat → Tools → Create images → Thinking, and selecting “Thinking with 3 Pro.” Pro tip: over‑explain instructions to give the model sufficient context.

Source