Google is upgrading its image generation model with new editing capabilities, higher resolutions, more accurate text rendering, and the ability to search the web. The new model is called Nano Banana Pro and is built on Google’s latest large language model, Gemini 3, which was released earlier this week.
The company states that Nano Banana Pro improves on its predecessor, Nano Banana, with the ability to create more detailed images and accurate text. It can also generate text in different styles, fonts, and languages. The model includes web searching capabilities, allowing users to perform tasks like looking up a recipe and generating flash cards.
Google says Nano Banana Pro is designed to give professionals more control over images. Users can control aspects such as camera angles, scene lighting, depth of field, focus, and color grading. While the previous model had a resolution cap of 1024 by 1024 pixels, users can now generate 2K or 4K images with Nano Banana Pro.
The company noted that although Nano Banana Pro generates higher quality images, it is slower and more expensive than the original model. The original model cost 0.039 dollars per 1024 pixel image. In comparison, the new model costs 0.139 dollars for each 1080p or 2K image and 0.24 dollars for every 4K image.
The new model can use six high-fidelity shots or blend up to 14 objects within an image. It can also maintain consistency and resemblance for up to five people. The company has released a demo app where users can try some of these capabilities.
Nano Banana Pro is being rolled out across many of Google’s existing AI tools. The Gemini app will now use the new model to generate images by default. However, users on the free subscription tier will be able to use the model for a limited number of images, after which they will be switched back to the original Nano Banana model.
Google AI Plus, Pro, and Ultra subscribers will receive higher generation thresholds, though the company did not disclose the exact limits. These subscribers will also get access to the model within Notebook LM.
Google is also making the model available in search through AI mode for AI Pro and Ultra subscribers in the United States. Ultra subscribers can access the model in the company’s video tool, Flow, and it is available to Workspace customers in Google Slides and Vids as well.
Developers can access Nano Banana Pro through the Gemini API, Google AI Studio, and the company’s new IDE, Antigravity.
The company is also integrating SynthID, its technology to watermark and detect AI-generated images, into the Gemini app. Users can upload an image, and the chatbot will tell them if the image has been created or modified by the company’s image models. Google did not mention if it also plans to support other AI-watermarking standards such as C2PA.