Google released the next-generation version of its image generation and editing artificial intelligence (AI) tool "Nanobanana," called "Nanobanana2." It features a significant boost in generation speed while maintaining the existing "Pro"-level high-quality performance.
Google said on the 26th (local time) it will roll out "Nanobanana2," based on "Gemini 3.1 Flash Image." It said the model combines the advanced reasoning of "Nanobanana Pro," unveiled in Nov. last year, with faster processing speed.
As a result, users can now use "Pro"-level high-quality images, previously a paid feature, for free at a faster speed than before.
It also supports translating and localizing text in images into other languages. By linking with real-time web search and Gemini's world knowledge, it can produce visual materials that reflect the latest information, making it highly useful for infographics and data visualization work.
It can maintain consistency for up to five characters and up to 14 object elements in a single task, reducing issues of changing character appearances when producing book illustrations or comics.
It supports resolutions from 4K ultra-high definition down to 512-pixel low resolution, and offers various aspect ratios such as 4:1 and 1:8, expanding creative options. For professionals, it also adds a function to adjust the model's level of reasoning, enabling more precise results in complex tasks.
Nanobanana2 will be installed as the default model in major Google products, including the Google Gemini app, starting today.
Meanwhile, since its first unveiling in Aug. last year, Nanobanana drew 13 million new users to the Gemini app in four days and generated more than 5 billion images by mid-Oct., marking a strong showing.