I Tested Google's Insane New Nano Banana Image AI

Google just released a brand new AI image model inside Gemini. It’s got a very strange name—Nano Banana. But the editor is incredibly good.

Specifically, while it’s just okay at creating images, it’s absurdly good at editing them.

Here’s a video showing some super cool examples from my testing, including my Grandpa’s old war photos coming to life and walking around (watch on YouTube):

Nano Banana can edit photos using text prompting. I asked it to make my dog Lance…

…dyed in rainbow colors, and it obliged.

The model also turned one of my grandpa’s 1940s war photos into a full color image:

These are fun examples. But from a practical perspective, this would be an amazing way to do lots of things for your business:

• Avoid hiring models by merging photos of your new product with photos of people, to create realistic images where they’re using the product. I demonstrate a version of this in my video above.

• Turning drawings or 3D renderings into realistic images to show how a building product or prototype would look once it’s completed.

• Showing clients how furniture would look in their space by realistically adding the furniture to a photo of their actual room.

• Editing a thumbnail image by merging multiple images together, to create a super compelling and realistic YouTube thumbnail. How do you think I made the thumbnail for the video above? :)