Google launches Nano Banana Pro, a next-generation image model powered by Gemini 3 Pro

Google has introduced Nano Banana Pro, the upgraded version of its popular Nano Banana image model, bringing a new level of precision, contextual understanding and creative depth. Built on Gemini 3 Pro, the model is designed to help users turn ideas, sketches and data into rich, accurate visuals in seconds, while also improving text rendering, multilingual output and complex scene composition.

The launch follows the success of Nano Banana, released only a few months ago, which helped casual creators restore old images, generate figurines and build quick edits. Google now says Nano Banana Pro is its most powerful image generation and editing tool for both everyday users and creative professionals.

The new model uses advanced reasoning and real-world knowledge to generate visuals that are more meaningful, accurate and context aware. It can interpret educational material, create detailed diagrams, produce recipe guides, and visualise real-time information by grounding itself in Google Search.

Users can expect better infographics, stronger visual explainers and image outputs that reflect up-to-date facts. For example, the model can generate weather-based graphics by pulling current conditions via Search, or produce data-rich educational illustrations.

Google says Nano Banana Pro is also capable of creating images with clean, legible text in a range of languages. This includes posters, calligraphy, product mock-ups and branding layouts. Thanks to Gemini 3’s multilingual reasoning, the model can translate and localise text inside images without breaking visual consistency.

Nano Banana Pro supports complex scene building by combining up to 14 input images while maintaining the identities of up to five people. This allows richer fashion editorials, product mock-ups, cinematic composites and even photorealistic 3D-style visuals.

Google has added studio-grade creative controls that let users modify lighting, adjust focus, change camera angles, switch day scenes to night, apply colour grading or recompose images for different aspect ratios. Output can be exported in 2K and 4K for print or social media.

Nano Banana Pro is rolling out across Google’s ecosystem:

    •    Gemini app: Available globally when selecting the ‘Thinking’ model. Free users get limited Pro quotas. Google AI Plus, Pro and Ultra subscribers get higher limits.
    •    Search: Available in AI Mode in the United States for Google AI Pro and Ultra users.
    •    NotebookLM: Rolling out globally for subscribers.
    •    Google Ads: Upgraded to Nano Banana Pro for advertisers.
    •    Google Workspace: Available in Google Slides and Google Vids.
    •    Developers: Launching through the Gemini API, Google AI Studio and Vertex AI for enterprise.
    •    Filmmakers and creatives: Rolling out to Google AI Ultra subscribers in Flow for precise frame control.

Google is also expanding its responsible AI measures. Every AI-generated image includes an imperceptible SynthID watermark. Users can now upload an image to the Gemini app and ask whether it was generated by Google AI. This will soon expand to audio and video.

A visible watermark will remain on images generated by free and Pro tier users, although Google AI Ultra subscribers and users of Google AI Studio will receive clean exports with no visible mark for professional use.