Google introduced Nano Banana Pro, a major upgrade to its artificial intelligence for image generation and editing. Announced on Thursday, this cutting-edge model leverages the company’s advanced Gemini 3 Pro large language model. By enhancing reasoning capabilities and integrating real-world knowledge, Nano Banana Pro empowers users to create highly detailed and accurate images
A key improvement in this release is the quality of the output. Users can now download high-resolution images generated on the Gemini platform in 2K and 4K definition. This is a substantial step up from previous versions.
Nano Banana Pro live on @runware – D0 launch! 🍌
Built on Google’s new Gemini 3 Pro, it offers:
– High-res image generation (2K & 4K)
– SOTA text rendering
– Precise editing controls
– 14 input images combined into 1 output image pic.twitter.com/ClvdeCaRH0
— Runware (@runware) November 20, 2025The Mountain View-based tech giant detailed the model’s availability in an official blog post. Nano Banana Pro will be accessible to all users, but with varying rate limits. Free users will have the lowest access quota. They will revert to the older Gemini 2.5 Pro-powered model once their limit is exhausted.
Paid subscribers of Google AI Plus, Pro, and Ultra tiers will receive significantly higher usage quotas. The model will also be integrated into AI Mode for paid subscribers in the US. It will also be available within NotebookLM for paying customers globally.
Meet Nano Banana Pro, on Replicate
Google's SOTA image generation and editing model. It has better world knowledge, text rendering, and precision controls. Plus 4K outputs. pic.twitter.com/aGgjT5mdLq
— Replicate (@replicate) November 20, 2025Google designed Nano Banana Pro to transform ideas into studio-quality visuals. It offers precise control and accurate multilingual text rendering. The model supports a wide array of professional applications through integrations with the Gemini app, Google Ads, Google Workspace, and the Gemini API.
Its advanced capabilities now include generating infographics, diagrams, and branded mock-ups by pulling live information such as weather and sports data. The model can render text directly within images in multiple languages. It maintains a consistent visual style across up to 14 input images.
Read: Google Announces $40 Billion Texas Expansion to Boost AI and Data Centers
For granular creative control, users can apply localised transformations, adjust camera angles, change aspect ratios, and fine-tune lighting and focus. The tool can produce social media and print-ready assets, generate storyboards, and convert handwritten notes into clean diagrams.
