Google rolls out Nano Banana 2 with 4K AI image generation

Google has officially launched Nano Banana 2, a major upgrade to its Gemini AI image generation tool, bringing 4K AI image generation, improved text rendering and better character consistency to free users.

Powered by the Gemini 3.1 Flash Image model, the update expands advanced AI image generator capabilities beyond paid subscribers.

Nano Banana 2 features

Nano Banana 2 is now rolling out across Gemini, Google Search, Google AI Studio and Google Ads, strengthening Google’s position in the competitive generative AI tools market.

Nano Banana 2's support for 4K resolution AI images doubles the previous 2K limit of the quality of the pictures. Users can now choose between 2K and 4K output, making their projects with high-quality visuals and professional content creation. The system uses Gemini’s real-world understanding to generate more precise and context-aware results.

Nano Banana 2 also addresses a common issue in AI-generated graphics: distorted or incorrect text. The updated Gemini AI image generator now produces clearer stylised text and supports accurate AI text translation within images. This makes it easier for businesses and creators to localise designs for different regions.

The new version improves character consistency in AI images, maintaining up to five characters across multiple scenes. It can also merge up to 14 objects from reference images into a single output. Complex instructions are interpreted more accurately, reducing missed details.