Google today unveiled the latest version of its popular “Nano Banana 2” image generation model. Technically called the Gemini 3.1 Flash Image, it can create more realistic images than its predecessor. The model will also become the default in the Gemini app’s “Fast”, “Thinking”, and “Pro” modes.
The company first released Nano Banana in August 2025, encouraging people, especially in countries like India, to generate millions of images on the Gemini app. In November, the company released Nano Banana Pro, which allows users to create more detailed and higher-quality images.
The new Nano Banana 2 retains some of the high-quality features of the Pro model, but produces images faster. The company says it can produce images with resolutions ranging from 512p to 4K, with a variety of aspect ratios.
The Nano Banana 2 can maintain the consistency of up to five characters and the accuracy of up to 14 objects in a single workflow for better storytelling. Google says users can also submit complex queries with detailed image generation nuances. Additionally, users can create media with brighter lighting, richer textures, and sharper details.
Upon release, Nano Banana 2 will become the default rendering model for all Gemini apps. The company is also making it the default rendering model for its video editing tool, “Flow.”
In the search space, Nano Banana 2 will become the default tool for displaying Google search results through Google Lens and in AI mode in 141 countries on the Google app and web on both desktop and mobile devices.
On Google’s higher-end plans, Google AI Pro and Ultra subscribers can continue to use Nano Banana Pro for specialized tasks by regenerating images via a three-dot menu.
For developers, Nano Banana 2 will be available for preview via the “Gemini API”, “Gemini CLI”, and “Vertex API”. It will also be available through AI Studio and the company’s Antigravity development tool, which was released last November.
The company said that all images created using the new model will have a SynthID watermark, Google’s mark for AI-generated images. The images also comply with the C2PA Content Credentials, an industry organization that includes companies like Adobe, Microsoft, Google, OpenAI, and Meta.
Google said that people have used the synthID verification feature more than 20 million times since it launched in the Gemini app in November.
Credits:
All images courtesy of Google.
