Google DeepMind’s Nano Banana Pro enables high-fidelity image generation and multilingual text rendering at up to 4K.
Gemini users on the free tier will get a limited number of Nano Banan Pro generations
Photo Credit: Google
Google released Nano Banana Pro, an improved version of the artificial intelligence (AI) editing model powered by Gemini 3 Pro, on Thursday. Calling it its state-of-the-art (SOTA) image generation and editing model, the Mountain View-based tech giant said Nano Banana Pro is capable of reasoning and accessing real-world knowledge to generate requested images. The company has also improved the quality of the output images, and it is now possible to download 2K and 4K images generated using the Gemini platform.
In a blog post, the tech giant announced and detailed the new version of the image editing model. The main change is that the tool is now powered by Gemini 3 Pro, rather than Gemini 2.5 Pro, bringing all the reasoning and contextual intelligence of Google's frontier large language model (LLM).
Coming to availability, Google said that Nano Banana Pro will be available to all users with varying rate limits. Free users will get the lowest rate limits, and once exhausted, they will revert to the Gemini 2.5 Pro-powered version of the model. Google AI Plus, Pro and Ultra subscribers receive higher quotas. The model will also be available in AI Mode, but only to users in the US for paid subscribers. It will also be available within NotebookLM for the paid users across the globe.
The company says the tool is designed to turn ideas into studio-quality visuals with improved control and accurate multilingual text rendering. It supports creatives, professionals and enterprises through integrations in the Gemini app, Google Ads, Google Workspace and the Gemini application programming interface (API).
It can now generate infographics and diagrams (or any other context-based visuals) as well as branded mock-ups by accessing live information such as weather and sports. The model supports text rendering directly in images, even in multiple languages, and applies consistent visual styling across up to 14 input images or representations of up to five people. Users can prompt it to generate a storyboard, convert handwritten notes into diagrams, or design high-fidelity visuals with studio-level controls.
In terms of granular creative control, Nano Banana Pro offers localised transformations, camera-angle changes, aspect-ratio shifts, lighting and focus adjustments. The model supports a wide range of outputs with up to 2K and 4K generations. Google says users can instruct the tool to generate social media or print-ready outputs.
Google embeds its digital watermark technology, SynthID, in all images generated by its tools, and provides a verification feature in the Gemini app to check whether an image was generated by Google AI. Visible watermarks are used for free and Pro users, while Ultra subscribers and enterprise tools will remove visible marks for professional work.
Catch the latest from the Consumer Electronics Show on Gadgets 360, at our CES 2026 hub.