Just when you thought Nano Banana couldn't get any better, they hit us with a sequel.
Hollywood has a rule: the sequel has to be bigger, faster, and more explosive than the original. Google has apparently been taking notes, because Nano Banana 2, AKA Gemini 3.1 Flash Image, was announced today and is ready for its close-up.
The original Nano Banana arrived last August to considerable fanfare, followed by Nano Banana Pro in November. Now we've got the second installment, and like any good sequel, it takes everything that worked in the first two and smashes them together. Pro's intelligence and world knowledge? In there. Flash's speed? Also in there. A runtime that won't have you staring at a loading spinner? That too. Unlike most movie franchises, this one actually seems to be getting better rather than quietly declining while the studio hopes nobody notices.
The headline feature is that Nano Banana 2 taps into Gemini's real-world knowledge base and live web search to ground its image generation in actual, current information. This means when you ask it to render something specific, it has real reference material to work from rather than just vibing and hoping for the best. It also means the model can help generate infographics, turn notes into diagrams, and produce data visualizations. The kind of stuff the original N.B. couldn't pull off nearly as well.
Text in images has always been AI generation's awkward villain origin story. While we've been super impressed by the progress Nano Banana has already made, we're now seeing another serious run at blowing our minds. Legible, accurate text for mockups and greeting cards, and the ability to translate text that already exists inside an image.
The subject consistency feature is where it really comes in handy for anyone doing narrative or design work. The model can maintain the appearance of up to 5 characters and 14 objects throughout a workflow. If you've ever watched an AI storyboard quietly change your main character's face between panels like some horror movie, you know exactly why this is worth celebrating. Resolution support now extends to 4K, so the production value of your outputs finally matches your ambition.
As for the theatrical release of this sequel, they're going wide. Nano Banana 2 is rolling out today in the Gemini app as the new default (Nano Banana Pro is still accessible for AI Pro and Ultra subscribers via the three-dot menu, like a director's cut for power users), in Google Search through AI Mode and Lens, in AI Studio and the Gemini API, in Vertex AI on Google Cloud, and in Flow at zero credits for all users: no limited engagement, no waiting for the streaming release.
Google is also continuing to build out its provenance tools, pairing SynthID watermarking with C2PA Content Credentials so you can tell not just whether an image was AI-generated, but how. SynthID verification has been used 20 million times since November. C2PA verification is coming to the Gemini app soon; consider it the post-credits scene.
We've been fans of the Nano Banana franchise from the start, and it looks like the sequel delivers. See for yourself and give it a try. Send in your Bugdroid creations to us, and we'll spotlight them on the Android Faithful podcast!
