Nano Banana 2.1: Google halves the price of AI images

On October 6, 2026, Google launched a new version of its AI image model, Nano Banana 2.1. According to several technical reports, it happened quietly, without a major blog post or published benchmarks (standard tests that measure quality).

The most important news for businesses is the price. Nano Banana 2.1 costs half as much per generated image compared with its predecessor, Nano Banana 2, at all resolutions where both are available.

What does it cost?

According to the developer site Apidog, the prices for developers who use the model via Google's API (the technical entry point where apps connect to the AI model) are:

- 1K image: about 0.034 dollars - 2K image: about 0.050 dollars - 4K image: about 0.076 dollars

By comparison, Nano Banana 2 cost around 0.067 dollars for a 1K image when it was released in February 2026.

There is a catch, though. According to Apidog, the price for the text instructions you send in (input) has risen from 0.50 to 1.50 dollars per million tokens. Tokens are small chunks of text that AI models calculate in. This means the savings may be smaller if you send very long instructions. There is also no free quota in the Gemini API, and the smallest resolution of 512 pixels has been discontinued.

What can the new model do?

Google highlights better visual design, more precise text in images, editing with masks (you mark an area and change only that) and greater consistency, so people and objects look the same across multiple edits. The images are described as more natural.

The model has been rolled out in Google's AI Mode in search, the video tool Flow, the developer tool Google AI Studio and the design tool Stitch.

What does it mean for you?

As an ordinary user, you are unlikely to notice the price directly. But when it becomes cheaper to make images, apps and services can become cheaper or offer more images. Companies that make many images can save money. Since Google has not published tests, you should test for yourself whether the quality holds up before switching.

What it means for you

If you or your company need many AI images, for example for ads, a website or social media, it may become cheaper if your tools use the new model. If you are an ordinary user, you are unlikely to notice the price directly, but apps may become cheaper or give you more images. If you are going to switch, first test some of your own images and check whether long text instructions eat up the savings. You can also try the model yourself in Google AI Studio or in AI Mode in Google Search.