Google today announced Nano Banana 2, the latest version of its popular image generation model. Technically a Gemini 3.1 Flash Image, this new model can create more realistic images than its predecessor. This model will also be the default for Fast, Thinking, and Pro modes in the Gemini app.
The company first launched Nano Banana in August 2025, and people started generating millions of images on the Gemini app, especially in countries like India. In November, the company released Nano Banana Pro, which can create more detailed, high-quality images.
The new Nano Banana 2 retains some of the high-fidelity characteristics of the Pro model, but produces images faster. The company says it can create images with resolutions from 512 pixels to 4K and a variety of aspect ratios.
Nano Banana 2 helps improve storytelling by maintaining consistency for up to 5 characters and fidelity for up to 14 objects in a single workflow. Google says users can also issue complex requests that include detailed nuances of image generation. Additionally, users can create media with more vibrant lighting, richer textures, and sharper details.
With this release, Nano Banana 2 will become the default model for image generation across all Gemini apps. The company has also made this the default model for image generation in its video editing tool Flow.
In search, Nano Banana 2 will be the default in Google search results through Google Lens and AI mode in 141 countries across Google apps and the web across desktop and mobile.
Google’s high-end plans, Google AI Pro and Ultra, allow subscribers to continue using Nano Banana Pro for specialized tasks by regenerating images from a three-dot menu.
tech crunch event
boston, massachusetts
|
June 9, 2026
For developers, Nano Banana 2 will be available in preview through the Gemini API, Gemini CLI, and Vertex API. It will also be available through AI Studio and the company’s development tool, Antigravity, released last November.
The company says all images created with the new model will include the SynthID watermark, Google’s mark that indicates images generated by AI. These images are also interoperable with C2PA Content Credentials, created by an industry group made up of companies such as Adobe, Microsoft, Google, OpenAI, and Meta. Google says people have used SynthID authentication more than 20 million times since it launched in the Gemini app in November.
