Google reveals improved Nano Banana Pro built on latest Gemini 3
Google DeepMind has rolled out Nano Banana Pro, an image model built on Gemini 3 Pro with improved text rendering, multi‑image consistency and fine‑grained controls, now reaching the Gemini app, Ads, Workspace and developer platforms with SynthID watermarking.
Google has unveiled Nano Banana Pro, a new image generation and editing model built on Gemini 3 Pro, promising greater control, improved text rendering and deeper real‑world knowledge for creating visuals across Google’s consumer, enterprise and developer tools.
Nano Banana Pro is a step up from the earlier Nano Banana (Gemini 2.5 Flash Image), adding stronger reasoning and “world knowledge” to help users turn prompts, notes and data into context‑rich infographics, diagrams and studio‑grade compositions.
It supports legible, correctly rendered text directly inside generated images, with multilingual output for localisation.
Key capabilities and controls
- Reasoning and grounding: The model draws on Gemini 3 Pro’s reasoning to produce accurate visuals and can connect to Google Search to ground creations in real‑time information such as weather or sports snapshots.
- Text inside images: It generates sharper, more reliable text for posters, mock‑ups and long paragraphs, with enhanced multilingual support.
- Complex compositions: Users are able to blend up to 14 input images while maintaining the resemblance of up to five people, and export in a range of aspect ratios at 2K and 4K resolutions.
- Fine‑grained edits: Localised editing allows selection, refinement and transformation of specific regions, with controls for camera angle, focus, colour grading and lighting (for example, day‑to‑night).
Availability across products
Google is rolling out Nano Banana Pro widely, offering a choice between the original Nano Banana for quick edits and Nano Banana Pro for complex work.
Consumers can see it in the Gemini app’s “Create images” with the Thinking model; free‑tier users receive limited quotas before reverting to Nano Banana, while Google AI Plus, Pro and Ultra subscribers have higher quotas.
In AI Mode in Search, Nano Banana Pro is available in the United States for Google AI Pro and Ultra subscribers.
More details on availability include:
- Creative and productivity tools: Image generation in Google Ads is upgraded to Nano Banana Pro; it has also started rolling out to Workspace customers in Slides and Vids.
- Developer and enterprise stack: The model has started rolling out via the Gemini API and Google AI Studio, is available in Google Antigravity for UX layouts and mock‑ups, and is available in Vertex AI for scaled creation, with support coming to Gemini Enterprise.
- For creatives: Google AI Ultra subscribers using Flow, Google’s AI filmmaking tool, have begun to gain more precise frame‑level control.
Provenance and labelling
All media produced by Google tools will reportedly continue to carry Google’s imperceptible SynthID watermark.
Google has also added a consumer check: people can upload an image to the Gemini app and ask whether it has been generated by Google AI.
Visible watermarks (the Gemini sparkle) remain on images from free and Google AI Pro tiers; they are removed for Google AI Ultra subscribers and in Google AI Studio. Google said it will extend SynthID‑based checks beyond images to audio and video.