Google has announced Nano Banana 2.1, an updated AI model for image generation and multi-turn editing. Replacing the Nano Banana 2 model released in early 2026, version 2.1 is powered by Gemini 3.6 Flash and acts as a lower-cost, higher-efficiency counterpart to Google’s flagship Nano Banana Pro model. The company announced that generation prices have been cut roughly in half, with 1K images dropping to 3.36 cents and 4K images priced at 7.56 cents.
According to Google, Nano Banana 2.1 brings technical upgrades in visual quality, text rendering, infographics, and wide panorama outputs. It can maintain visual consistency across up to four characters and ten objects by processing up to 14 reference images simultaneously. The model also features integrated access to Google Search and adjustable reasoning compute levels (minimal, medium, and high) that influence output fidelity.
The model is currently rolling out across Google’s developer and consumer ecosystem, including Google AI Studio, the Gemini app, Search AI Mode, Google Ads, and Gemini Enterprise. Google noted it will deprecate the previous generation model, gemini-3.1-flash-image, on October 29, 2026.
Why it matters
Cuts API media generation costs by roughly 50%, significantly improving unit economics for enterprise image generation workflows.
Enhances multi-turn asset generation by maintaining visual consistency across up to 14 reference images and multiple subjects.
Integrates native web search and controllable reasoning compute directly into mid-tier media production models.
Source: the-decoder.com



