Google has announced the launch of Nano Banana 2, its new image generation model powered by Gemini 3.1 Flash Image. According to the company’s statement, the new model combines the reasoning capabilities, world knowledge and performance of the well-known Pro model, but does so at “lightning” speed. The upgrade in processing speed is intended to allow users to perform quick edits and create a large number of variations and outputs based on a single command (Prompt) in a short time.
The new model is expected to make capabilities that were previously reserved exclusively for Pro subscribers accessible to a wider audience of users. Among other things, the model is able to pull information and images in real time from web searches in order to generate complex outputs such as infographics and diagrams. In addition, the ability to integrate text within images has been improved, which will make it a relevant tool for creating marketing materials and greeting cards.
One of the key innovations in Nano Banana 2 is the ability to maintain visual consistency of characters (Character Resemblance) for up to five different characters within a single workflow. This is a critical capability for storyboard creators and visual storytelling. The model can follow precise instructions for complex requests and is capable of producing outputs in 4K resolution, including richer textures and sharper details compared to previous generations.
The previous Pro model in the series already demonstrated a very high level of realism, to the point that it was difficult to distinguish that the result was generated by artificial intelligence, which led Google to restrict its use due to demand overload. Now, Nano Banana 2 will replace the Pro model in the Gemini app and become the default model in Search (AI Mode), in Lens and in the Flow creative studio. However, Google AI Pro and Ultra subscribers will continue to have access to the Pro version for specific specialized tasks.