What is Nano Banana?
Nano Banana is the unofficial nickname the AI community gave Google's Gemini image generation model. The name stuck because of an Easter-egg internal codename. Today it refers to the same family of models powering Gemini's image features.
Where the name came from
The "Nano Banana" name leaked from internal Google testing in 2025. It was never the official product name, but the community adopted it because:
- It was easier to say than "Gemini image generation".
- It distinguished the image-specific capability from the broader Gemini chat model.
- The name went viral on X and Reddit before Google could rename.
What Nano Banana actually is
Technically: a multimodal image generation model in Google's Gemini family. It can:
- Text-to-image: generate from a written prompt.
- Image-to-image (edit): modify an uploaded image while preserving identity.
- Multi-image composition: combine two reference images into a single output.
- Style preservation: keep a character or product consistent across multiple generations.
Why it matters
Three years ago, image AI was either Stable Diffusion (open, but local) or DALL·E / Midjourney (closed, paid). Nano Banana brought Google-quality output to a public API at a price point that lets web apps offer it cheaply.
Noaroo runs directly on top of this model family — sh-flash for fast drafts, sh-pro for photoreal finals.
How to try it
- Open the [playground](/) — no signup needed to browse.
- Sign in with Google — 3 free credits applied automatically.
- Type a prompt. Pick flash or pro. Generate.
That's it. No API keys, no Cloud Console, no quota requests.
FAQ
Is Nano Banana the same as Gemini?
Image generation inside Gemini is what people call Nano Banana. Same model, different name.
Did Google officially rename it?
Google calls it the Gemini image model. The community name stuck regardless.
Where can I use Nano Banana?
Through Google AI Studio (developer-focused), or through wrappers like Noaroo that handle UI, history, billing, and edit mode.