With the spread of artificial intelligence, the technology to create images and videos at will is at hand. Anyone can create the desired image-video by writing appropriate prompts. To make this task easier and faster, tech giant Google has further developed their ‘Nano Banana’ model. The new model is named ‘Nano Banana 2’.
TechCrunch, a California-based technology news outlet, reported this information on Thursday. Google announced this new model today.
According to TechCrunch, ‘Nano Banana 2’ is technically known as Gemini 3.1 Flash Image. Compared to the previous version, this new model is capable of producing more realistic images. It will also be used as the default image generation model in all Gemini app’s Fast, Thinking and Pro modes.
Google released the first Nano Banana model in August 2025. Since then, users around the world have created billions of images using the Gemini app. Especially in India this app is used the most. Later in November of that year, Google introduced the ‘Nano Banana Pro’, which allowed for higher quality images.
The new ‘Nano Banana 2’ can produce images faster while retaining some of the high quality features of the previous ‘Pro’ model. Google claims, through this model, from 512 pixels to four-K resolution, images of different proportions can be created.
The model can maintain up to five character sequences simultaneously to facilitate storytelling or story-based content creation. It can hold up to 14 objects accurately in a single workflow. Users can also give complex and precise instructions, based on which more perfect images will be created. Also, media content can be created with better lighting, richer textures and sharper details.
With the introduction of this new model, the default image generation model in all Gemini apps will be ‘Nano Banana 2’. It will also be used as the default image model in Google’s video editing tool Flow.
Meanwhile, Google search is also changing. Through Google Lens and AI mode, ‘Nano Banana 2’ will serve as the default model in Google App and web search results on both desktop and mobile platforms in 141 countries.
However, users of Google’s higher subscription plans – Google AI Pro and Ultra – can also use Nano Banana Pro from the three-dot menu.
For developers, ‘Nano Banana 2’ is available in preview version via Gemini API, Gemini CLI and Vertex API. In addition, it can be used in AI Studio and Google’s development tool Antigravity, released last November.
According to Google, all images created with this new model will have the SynthID watermark. This is Google’s own system for identifying AI-generated images. Also, these images are compliant with C2PA Content Credentials standards. According to Google, SynthID verification has been used more than 20 million times since the Gemini app launched last November.
