Nano-banana, the mysterious image generation model that had appeared on LMAreana a few days ago and had wowed people with its editing capabilities, might finally have a claimant.
Google has appeared to hint that the viral nano-banana image creation AI model is from its stable. Logan Kilpatrick, head of Google AI Studio and the face of the company on X, has posted a single banana emoji on X.
This almost certainly points to the fact that it is indeed Google which is behind the model. There had been speculation online that nano-banana was a Gemini model, and Kilpatrick’s post all but confirms it.
Nano banana had quietly appeared on LMArena a few days ago. Users had been impressed with its prompt adherence and its speed. Nano banana had both text to image and image to image capabilities, and was especially adept at preserving consistent characters or products.
“Forget spending hours staging shots. I gave Nano Banana 1 image and it built an entire sequence like it’s nothing. Speed + consistency + vibes = next level,” an X user had written.
Professor Ethan Mollick had also tested the model, and said that “prompt adherence has become incredibly solid across multiple models”.
Another X user had called the nano-banana model ‘impressive’.
Another user had called it the best image editing model yet.
It now appears that it is indeed Google behind the nano-banana model. Google already has some competent image generation models, but they haven’t quite stood out from the competition like its Veo models have for video. With nano-banana, Google appears to have created a state-of-the-art image generation model that seems to have a leap in capabilities over existing approaches. It remains to be seen when Google officially launches the model, and with what sort of pricing, but given the buzz around the model even before its official release, Google might already have a hit on its hands.