Hands-On With Nano Banana 2, the Latest Version of Google’s AI Image Generator

Google has officially released Nano Banana 2, the latest iteration of its AI image generator, which now serves as the default model within the Gemini chatbot. This update merges the established capabilities of the previous Nano Banana Pro—notably its proficiency with text rendering and integrated web searching—with a substantial boost in generation speed. Accessible via the Gemini app, website, and Google's suite of services like Search, AI Studio, and Cloud, the tool is positioned as a more powerful photo editor. However, initial hands-on testing reveals persistent challenges, such as generating unconvincing images and pulling incorrect real-time data, as evidenced when it created a weather infographic with outdated forecasts.

The technical advancement of Nano Banana 2 lies in its faster processing and its ability to manipulate existing images with impressive detail, faithfully recreating elements like clothing patterns and jewelry from source photos. Despite these improvements, the model struggles with nuanced requests, often producing results that are more comical than intended, such as making a person look elderly instead of just pruny from a hot tub. Its integration across Google's ecosystem and its real-time data pull for creating infographics point to its utility, but also serve as a reminder of the need for critical scrutiny of AI-generated content. The tool's evolution underscores the rapid progress in photorealistic AI image manipulation and its growing accessibility to the public.