Google's latest AI model has sparked another viral transformation trend, with millions of users turning themselves into miniature Bandai-style figurines that look pulled from a Tokyo toy store shelf.

The tool, nicknamed Nano Banana after its internal codename at Google DeepMind, officially launched as Gemini 2.5 Flash Image in late August. Within two weeks, users generated over 200 million images. The Philippines alone accounts for 25.5 million image creations just days after its launch, making it the top region globally for use.

"From photo to figurine style in just one prompt," Google's Gemini team posted on X September 1, kicking off tutorials that spread the trend further. Josh Woodward, Google's VP, reported 10 million new Gemini users joined specifically for the feature.

The digital figurines appear as 1/7 scale collectibles on clear acrylic bases, often displayed next to packaging boxes and computer screens showing 3D modeling software. The AI captures facial features, clothing details, and poses with enough accuracy that results frequently fool viewers into thinking they're real product photos.

This marks the third major AI portrait trend of 2025. ChatGPT's GPT-4o sparked the Studio Ghibli wave in March, with users transforming photos into soft, anime-style portraits reminiscent of Hayao Miyazaki's films. Sam Altman changed his X profile picture to a "Ghiblified" version of himself as servers struggled under "biblical demand."

April brought the Barbie Box Challenge through ChatGPT, where people became plastic dolls in blister packaging, complete with accessories like tiny laptops and coffee mugs. LinkedIn professionals embraced it for personal branding, turning themselves into "executive action figures."

Nano Banana isn't the only player in this space. Alibaba's Seedream4, released just a few days after Google’s model, produces outputs with comparable quality and sometimes better consistency in complex poses. For those preferring open-source alternatives, Flux Kontext offers the most powerful option for local deployment. It integrates into complex workflows and remains the only model without content restrictions, though that flexibility comes with the usual responsibilities of self-hosted AI.

How to turn yourself into a figurine, for free

Creating your own Nano Banana figurine takes under a minute. Visit gemini.google.com or open the Gemini app—the basic version is completely free, though your daily generations are limited. 

Click on the option to try Nano Banana, which you can find it in the banner on top. It is also under “tools” as “Create Images.” (Google is showing banana emojis everywhere, so you won’t have any trouble finding it.)



Next, upload your photo or the photo of the person you want to turn into a doll, preferably a full-body photo with good lighting. You can do that by clicking on the “Plus” button next to “tools.” We trust you will be responsible, ok?

Once the image has been uploaded, paste in this prompt:

"Create a 1/7 scale commercialized figurine of the character 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."

There are some tweaks you can make to customize your doll.  Specifying materials like "PVC figure" or "polished resin" increases realism. Dynamic poses work better than static standing positions—arms outstretched or mid-action yield more convincing figures.

If you don’t think the model is realistic enough, adding something like “The face is exactly the same” or “the face remains unaltered,” does the trick. But beware, you may end up with something too realistic and not really doll-like.

Don’t go too crazy, though. The free tier offers limited daily edits, while the pro version removes restrictions. Some platforms, like Freepik, give users unlimited iterations (and unlimited Wan 2.2 videos too) with the tradeoff of a small degradation in quality after a very generous threshold is met.

Advanced users chain multiple prompts for complex scenes. After generating the base figurine, you can improve your doll with further iterations—asking it, for example, to change the clothes, add props, alter the pose, etc. Nano Banana’s outstanding character reference makes it a very powerful tool that won’t degrade your face too much after many iterations. This is something that previous models used to struggle with.

The figurine aesthetic has proven particularly sticky on social media. Politicians in India and the Philippines posted their miniature versions, while TikTokers use the hashtag #NanoBanana to share increasingly elaborate scenes with multiple figures and custom dioramas.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。