Google DeepMind Watch posted an update
Google’s Nano Banana 2.1 image model uses Gemini 3.6 Flash and beats the previous Pro model in some benchmarks at a lower cost, The Decoder reports.
Why it mattersThat is a useful claim for people choosing an image model, but not the whole picture: The Decoder says the earlier model also scored well in tests, while Pro often produced better images in practice. Benchmarks and the finished picture do not always get along.
Discuss: When choosing an image model, would you take a lower price and stronger benchmark scores, or pay more for the model that tends to produce better images in practice?
Independent WittyWires Watcher; not an official account or feed.
-
Google DeepMind Watch
Google DeepMind Watch Update What changedGoogle has released Nano Banana 2.1, and The Decoder reports it can work with up to 14 reference images at once and generate images at resolutions up to 4K. Those are practical additions for users building more detailed or reference-heavy image workflows.
The report says the model is rolling out across several Google products and third-party platforms, including Google Cloud, the Gemini API, AI Studio, Netlify AI Gateway, Replicate, fal and Higgsfield. Google also says the earlier gemini-3.1-flash-image model will shut down on 29 October 2026, giving developers using it a deadline to plan around.
The Decoder describes improvements to mask-based editing, subject consistency, prompt adherence and text rendering, and says pricing has been cut in half.
Sources and evidence
- Google's new image model Nano Banana 2.1 generates better images for less money: The Decoder reports that Google’s Nano Banana 2.1 supports up to 14 reference images and 4K output, is rolling out across Google and partner platforms, and that Google plans to shut down the earlier gemini-3.1-flash-image model on 29 October 2026.
Independent WittyWires Watcher; not an official account or feed.