a puzzling Nano Banana AI image creator gained immense popularity, not solely due to its intriguing title, but rather because of the AI-generated visuals users continuously shared online. Some speculated that Nano Banana, while in development on LM Arena, was an unreleased model from Google that would soon be revealed. By the end of August, Google launched Nano Banana, disclosing its official designation, Gemini 2.5 Flash Image, and rendering it broadly accessible to users. Despite this, the public persisted in referring to it as Nano Banana, which Google ultimately embraced, incorporating the name across all platforms where the AI image generation model is presented.
Until recently, Google had not clarified the significance of Nano Banana, a label that seemed nonsensical in the overarching context of Google AI initiatives, where Gemini serves as the primary brand. Some might have speculated that Nano Banana was merely a random temporary label Google opted for when submitting the AI image generation model to LM Arena. It turns out that was precisely the case. Google confirmed that the designation was randomly selected in the middle of the night, not stemming from a meticulously devised plan. The name became popular as the model gained traction and users continued to refer to it as Nano Banana. In hindsight, the choice of name was a moment of late-night ingenuity, as it provided the AI model with a distinct identity.
[embedded content]
Gemini’s group product manager, David Sharon, appeared on the “Made By Google Podcast”, discussing Nano Banana, its development stage, the viral tests on LM Arena, its safety, and
