Google rolls out Nano Banana 2.1 across AI Studio, API and Gemini

On October 6, 2026, TestingCatalog reported on X that Google had begun rolling out Nano Banana 2.1, its image generation and editing model, across Google AI Studio, the APIs, and Gemini. The post quoted Google's own description of the model, which it says delivers enhanced factuality, visual fidelity, and multi-turn consistency at Flash scale. Image output is priced at $30 per 1,000,000 tokens.

A day later, on October 7, cnBeta reported that Google had formally launched Nano Banana 2.1 as an upgrade to Nano Banana 2. According to that report, the new version focuses on stronger image editing, visual design, text rendering, and subject consistency, and has started arriving in Gemini as well as in several other Google AI products aimed at both consumer and enterprise users.

The two accounts cover the same model but frame it differently. The first treats the news as a rollout in progress and leads with the official capability language and the token price; the second treats it as an official launch and lists a somewhat different set of strengthened abilities. Neither report says when the rollout will be complete, and neither provides a full list of the products that will receive the model. TestingCatalog's post also describes the model as sitting at Flash scale, a phrasing cnBeta does not use, while cnBeta's account of which capabilities improved is broader than the three points Google is quoted on. What both agree on is the name, the fact that the model covers both image generation and editing, and its arrival in Google AI Studio, the APIs, and Gemini. As of the latest report, Nano Banana 2.1 is available or arriving in those surfaces, with image output priced at $30 per 1,000,000 tokens.