Google has introduced an updated model for image generation, Nano Banana Pro, which operates on the latest large language model Gemini 3. Users can now create more detailed images in 2K and 4K resolution, whereas the previous version of Nano Banana supported a maximum of 1024 x 1024 pixels. The model allows editing images, changing the camera angle, lighting, depth of field, focus, and color scheme.
Nano Banana Pro generates text on images more accurately, supports various styles, fonts, and languages. A “search the internet” feature has been added, allowing users to, for example, find a recipe and create flashcards based on the information found. The model can combine up to 14 objects in one image and maintain similarity and consistency for up to five people across different images.
The cost of creating one image in Nano Banana Pro is $0.139 for 1080p or 2K and $0.24 for 4K. It is more expensive and slower compared to its predecessor, but users receive higher quality results. The service also offers a demo app to explore the model’s capabilities.
Nano Banana Pro is already integrated into the Gemini app, where free users can create a limited number of images, after which the system reverts to the basic Nano Banana. Paid subscribers of Google AI Plus, Pro, and Ultra receive extended generation limits and access to the model in NotebookLM, Google Slides, Vids, and the video tool Flow.
Developers can use Nano Banana Pro through the Gemini API, Google AI Studio, and the new Antigravity development environment. The Gemini app now includes a feature for checking images for signs of generation using SynthID, and in the future, Google plans to implement C2PA support for content authenticity verification.
