What Nano Banana is and how it works

Nano Banana is not a separate service, but a nickname “stuck” to Gemini 2.5 Flash Image, a new image generation model from Google.

The story behind the name started with the LMArena platform, where users test different neural networks blindly and vote for the best results. In August 2025, an unknown “nano-banana” model appeared there and quickly emerged as a leader in the photorealistic generation category.

A couple of weeks later, Google unveiled Gemini 2.5 Flash Image, and it became clear that this was the technology behind the mysterious name. Although the company had not originally planned to use the name Nano Banana, the internet quickly latched on to the model.

Give the selfies and group pics that belong in a museum the special treatment with our new nano banana image creation & editing update. pic.twitter.com/eTZGpIV08j

– Google Gemini App (@GeminiApp) August 29, 2025

The main advantage of the neural network is preserving portrait likeness and fine details. It carefully edits images and, if asked to change the hat in a photo, the model will change only it, leaving the rest of the elements without distortion. This makes Nano Banana closer to Photoshop than to classic generative models.

The tool works through the Gemini chatbot and Google AI Studio. The free version allows you to create or edit up to 100 images per day, which is enough for most scenarios.

Nano Banana’s interface is identical to familiar chatbots, and authorization is done through a Google account. It’s important to make sure the 2.5 Flash model is active, otherwise results may vary.

How to Make Your Own Figurine: Step-by-Step Instructions

The trend for figures from Nano Banana became popular because the result looks as realistic as possible and resembles expensive collector’s editions. Below — instructions on how to create such an effect yourself.

Step 1: Prepare a photo

Upload your image to Gemini chatbot. It can be a portrait, a full-length photo, selfies or even a shot of your favorite pet. The clearer the source, the more details Nano Banana will preserve. In our case, the “pioneer” was a dog, Barney Versetti, who was happy to test the new technology.

Step 2: Use a special promt

Next, insert a promt into the dialog with the chatbot, which sets the scene and attributes of the future figure. It is the one responsible for verisimilitude, including the stand, box and 3D model on the screen. It was, by the way, shared by the Google team itself, which makes the wording one of the most optimal in terms of content.

The original promt:

“Create a 1/7 scale commercialized figurine of the characters in the picture, in a realistic style, in a real environment. The figurine is placed on a computer desk. The figurine has a round transparent acrylic base, with no text on the base. The content on the computer screen is a 3D modeling process of this figurine. Next to the computer screen is a toy packaging box, designed in a style reminiscent of high-quality collectible figures, printed with original artwork. The packaging features two-dimensional flat illustrations”.

Step 3. Complete with details if desired

Users often modify the promt to add customized elements. For example, you can specify a profession or interest:

“Create a 1/6 scale commercialized figure of the character in the illustration, in a realistic style and environment. Place the figure on a clear acrylic stand. Next to the computer screen, display the ZBrush modeling process of the figure. Next to the computer screen, place a toy packaging box with the title [“Alorar Flasion”] printed with the original artwork”.

This way, you’ll get a figure with a personalized box with your profession or role printed on the packaging — be it “UX DESIGNER”, “GAMER” or any other.

Step 4: Get the result

Nano Banana uses the uploaded photo as a reference and integrates the character into a given scene. The result is an image that looks like a photo of a real collectible figurine on your desktop.

That’s what our Barney Versetti came out with:

The effect is appreciated for the combination of detail:

• the gloss and reflections of the figurine’s surface;
• clear acrylic stand;
• Bandai or Good Smile Company style box;
• a screen with a 3D model as a “proof” of production.

These elements make the picture feel real and are suitable for personal branding, viral content or just plain fun. Here, by the way, is another example of such an image with Incrypted CEO Ivan Pavlovsky:

Other Nano Banana Features

While the figurine trend has become a major infomercial, Nano Banana has much more functionality:

• Part editing. The model allows you to change individual elements in a photo, from hair color to headgear. Unlike many competitors, it does not touch unnecessary parts of the image;
• Fitting clothes and accessories. The service is able to “put on” a person specific clothes or glasses, and the result looks as if the thing was really present in the frame;
• Interior design. Nano Banana helps to add furniture or reorganize the room. It is important to specify what should remain unchanged (such as windows or flooring) to get a realistic result;
• Product photos. The model is used to generate advertising images. It can remove the background, place the item on a stand or package it in a gift box;
• Collages. It supports combining several images in one. For example, you can combine a portrait, architecture, and additional subjects;
• Restoration and colorization. Nano Banana restores damaged photos and adds color to black-and-white photos while maintaining portrait likeness.

These features make the tool versatile and allow it to be used both for entertainment on social networks and in practical tasks — online shopping, marketing and editing old family albums.

In general, we can say that Nano Banana is an example of how fast image editors based on artificial intelligence are developing. On the one hand, it а new tools for business, on the other hand, it launches viral trends like realistic personalized figurines. And it does it at a very decent level.