So, I finally got around to testing out the brand new Gemini Banana Nano image generation model from Google (docs here: link), if not for its performances just for the name!

This little model is designed to make high-quality image generation faster and lighter, and I couldn’t resist giving it a spin. My first experiment? Updating my profile picture.

For years, my main profile image was actually my wedding photo—and let’s just say that was about 15 years (and a few gray hairs) ago 😅. Time to bring things up to date!

I prompted Banana Nano using a recent photo of me taken just a few weeks ago at another wedding (not mine 🙂). The original shot wasn’t professional at all—messy lighting, other people in the frame—but the model managed to clean it up and generate a version that felt realistic, recognizable, and much more “me” right now.

But what’s even more exciting is how accessible this model feels. Instead of requiring huge infrastructure or endless GPU time, it’s designed to run faster and more efficiently—making it a great option for quick experiments, side projects, or even embedding image generation into lightweight apps.

Some quick impressions:

• ⚡ Speed: Results came back in seconds, not minutes.
• 🎨 Quality: While it’s lighter than full-size models, the outputs are sharp and coherent.
• 📱 Use cases: Perfect for mobile or browser-based tools where big models are overkill.

👉 I’ll share more tests soon, maybe even some side-by-side comparisons with other image models. For now, big thumbs up to Gemini Banana Nano for quick, lightweight, and shockingly accurate image generation.

And yes… after 15 years, I finally have a profile picture that looks like me today! 😁