Google made Nano Banana 2.1 generally available on 6 October 2026. The Gemini API changelog lists it as gemini-nano-banana-2.1, an update to Nano Banana 2, and marks the older gemini-3.1-flash-image as deprecated with no shutdown date set yet. Google DeepMind's model card, published the same day, says the model is built on Gemini 3.6 Flash and is shipping across the Gemini app, AI Studio, the Gemini API, AI Mode in Search, Google Ads, Google Flow and Google Stitch.
The listed changes are the ones art directors fight with every day: better visual quality, tighter prompt adherence, steadier characters across multi-turn edits and cleaner text. The model page adds multi-image fusion from up to 14 references, with up to four characters and ten objects, output at 1K, 2K and 4K, and a fix for tiling artefacts on the wide 1:4, 4:1, 1:8 and 8:1 ratios at 2K and 4K. Thinking runs from minimal to high, and answers can be grounded in Google Web and Image Search. On Google's own preference tests, the thinking version scores 1106 Elo on multi-character consistency against 978 for Nano Banana 2. API pricing drops to $0.0336 per 1K image and $0.0504 per 2K, roughly half the $0.067 and $0.101 Nano Banana 2 charges.
