Google reveals Nano Banana 2 AI image model, coming to Gemini today (arstechnica.com)

Google has announced the launch of Nano Banana 2, also known as Gemini 3.1 Flash Image, which becomes available today. This updated AI image generator builds on the original Nano Banana models, delivering enhancements that combine the high-quality output of the Pro version with the speed of the non-Pro variant. Notably, Nano Banana 2 leverages advanced knowledge from Google’s Gemini 3.1 language model, enabling it to produce more accurate image rendering and improved infographic creation, while achieving text accuracy in image outputs akin to Pro-level versions. For the AI/ML community, this release is significant as it marks a continued evolution of Google's image generation capabilities, aimed at enhancing visual storytelling through richer textures and vibrant lighting. The model can handle up to five characters and accurately render up to 14 different objects simultaneously, expanding the creative possibilities for users. Additionally, Google is broadening aspect ratios and resolutions, from 512px square to 4K widescreen formats. With this strategic push, Nano Banana 2 will replace all previous versions across various platforms, indicating Google’s confidence in its new AI model and setting a benchmark for future developments in AI-generated imagery.