Google Launches 'Nano Banana 2' for AI Art
Google just launched Nano Banana 2 as part of its Freepik AI Suite. The tool offers unlimited 1K and 2K image generations, giving AI visual creators and developers a powerful new resource for generative AI portfolio projects.
Nano Banana 2 is the public name for Google's third-generation AI image model, technically known as Gemini 3.1 Flash Image. The "Nano Banana" moniker originated from its codename on the LMArena crowdsourced evaluation platform in August 2025, where it quickly topped the text-to-image rankings before Google officially acknowledged it. The new model merges the high-speed performance of Gemini Flash with the advanced features previously exclusive to Nano Banana Pro. This allows for faster iterations and editing while improving creative control. For developers working on complex visual projects, the model can now maintain the consistency of up to five distinct characters and 14 objects within a single workflow. Within the Freepik AI Suite, Nano Banana 2 is one of several integrated models, sitting alongside video generation tools like Google's Veo and OpenAI's Sora. The suite functions as an all-in-one creative platform, offering AI-powered editing tools for background removal, upscaling, and a collaborative canvas called Spaces for managing project workflows. The integration is part of a wider rollout across Google's ecosystem. Nano Banana 2 is also being deployed in the Gemini app, Google Search, and Flow, Google's AI-powered video tool, indicating a strategy to embed this technology across its core products. This launch places Google in direct competition with a growing field of AI art generators relevant to developers, including tools with dedicated APIs like Prodia and open-source models such as Stable Diffusion. The availability of unlimited 1K and 2K generations on Freepik for subscribers is a strategic move to attract high-volume creators and developers to its platform. As AI-generated media becomes more widespread, Google is also emphasizing content provenance. The company is coupling its SynthID watermarking technology with C2PA Content Credentials to provide more context on how AI was used to create an image, a feature that has been used over 20 million times in the Gemini app since its launch.
