Google DeepMind published the model card for Nano Banana 2.1 on October 6, 2026.
What to know
- The card describes it as a member of the Gemini 3 series, based on Gemini 3.6 Flash.
- It accepts text and images as input, with a context window of up to 1 million tokens, and produces images and text. The card lists image output with a 4K token output and text output up to 64K tokens.
- The card lists these channels: the Gemini app, Google AI Studio, the Gemini API, Google Search AI Mode, Google Ads, Google Flow and Google Stitch.
- Google’s evaluation notes cover text-to-image, text rendering in several languages, editing, multi-image and multi-turn work.
Where it sits
Nano Banana 2 launched on February 26, 2026. The model card does not say Nano Banana 2 is retired, so check each Google product for which version it uses. See the Nano Banana review and the model tracker.
Source: Google DeepMind model card. We summarise in our own words; check the source for the full details.