How Nano Banana took over the internet and shot Gemini to the top of the app charts
Ever wondered why DeepMind’s viral new image model is called “Nano Banana”? We spoke with DeepMind’s Naina Raisinghani (Product Manager) and Philipp Lippe (Senior Research Scientist) to get the real story behind Nano Banana’s enormous success.
Speed is everything. Nano Banana generates high-quality images in 2-6 seconds — faster than the typical 20–60 seconds for most models. According to Lippe, waiting a minute for an image “breaks down the whole interaction, and user satisfaction definitely goes down.” Algorithmic improvements make this leap possible.
Consistency unlocks creativity. The model has generated over 5B images in less than a month and keeps facial features, expressions, and character details consistent across scenarios. Users can reimagine themselves, family, or pets in new contexts — taking a “family trip to Paris without actually having been there” or seeing “what you would look like on a 90s sitcom.” People also use it for personal photo shoots and restoring old family photos.
Reasoning enables smarter edits. Unlike previous models, Nano Banana can reason about input images. Users can upload a Google Maps screenshot and ask it to “imagine this scene from the point of view of a person standing there,” and it generates that perspective. Gemini’s “world knowledge” opens use cases that were previously impossible.
People are using it for work, too. The action figure trend captured social media, but professional applications gained traction as well. Developers use it for virtual try-ons, interior design tools, and advertising workflows. As Raisinghani puts it, it’s becoming “your stylist, your interior decorator, and your collaborator.”
So why is it called “Nano Banana“? According to Raisinghani, the name was chosen just for fun because it perfectly captures the model’s playful creativity.
via Superhuman