Google has released a new model for generating and editing images, Nano Banana 2.1. It is based on Gemini 3.6 Flash, replaces February’s Nano Banana 2 and offers improved quality of work with visual content at a significantly lower price.
Google has released a new model for generating and editing images, Nano Banana 2.1. It is based on Gemini 3.6 Flash, replaces February’s Nano Banana 2 and offers improved quality of work with visual content at a significantly lower price.
According to the released Google model card, the updated tool has received higher text reproduction accuracy, better character consistency and the ability to process up to 14 reference images simultaneously. The cost of generating an image with a resolution of 1K has almost halved — from $0.0670 to $0.0336, and a 4K image now costs $0.0756 instead of $0.1510. For comparison, the flagship Nano Banana Pro will cost $0.1340 for a 1K image.
Nano Banana 2.1 supports a context window of up to 1 million tokens at the input. The developers note that the model is able to maintain visual consistency for four characters and ten objects within a single session or dialogue. In addition, the generator connects to Google Search and allows you to choose one of three levels of reasoning depth (low, medium and high), which affect the detail of the final result.
Google describes the new model as a “more efficient analogue” of the Pro version, although it acknowledges its limitations. In particular, the model can still make mistakes with spatial localization of objects, and small text on 1K images sometimes remains blurry. The model is already available through the Gemini API, Google AI Studio, the Gemini application, and other ecosystem tools.
The previous version of the Nano Banana 2 was released in February 2026. In the summer, the company introduced the Nano Banana 2 Lite. Google emphasized that this version is much faster and more affordable than the previous one.
