s/demishassabisIMAGE MODELS•Oct 7
143
votes
7.2k
seen
Nano Banana 2.1 costs about 50% less than Nano Banana 2
Google has released Nano Banana 2.1 across the Gemini app, AI Studio, API, Google Flow, and third-party platforms. The image generation and editing model focuses on cleaner visual design, mask-based edits, better subject consistency, and more natural-looking results.
It supports 4K output, multiple reference images, and Google Search grounding. Early tests found stronger removal edits and better fidelity across repeated reconstructions, but some users reported weaker identity continuity than Nano Banana 2.0 in difficult editing chains. The model also showed up in Flow under the codename “beluga” before the wider announcement.
Techmeme reported that it is based on Gemini 3.6 Flash and costs about 50% less than Nano Banana 2.
It supports 4K output, multiple reference images, and Google Search grounding. Early tests found stronger removal edits and better fidelity across repeated reconstructions, but some users reported weaker identity continuity than Nano Banana 2.0 in difficult editing chains. The model also showed up in Flow under the codename “beluga” before the wider announcement.
Techmeme reported that it is based on Gemini 3.6 Flash and costs about 50% less than Nano Banana 2.
Timeline3
Oct 6
Google announced Nano Banana 2.1 with improvements to visual design, mask-based editing, and subject consistency.
Oct 6
Early testing cited 4K output, multiple reference images, Google Search grounding, and improved text rendering.
Oct 6
Techmeme reported that Nano Banana 2.1 is based on Gemini 3.6 Flash and priced about 50% below Nano Banana 2.
1 comment
Oct 7