
Nano Banana (aka Gemini 2.5 Flash Image) is Google/DeepMind’s new image generation and editing model.
This model is now in the Gemini app and API.
It can create entirely new images, tweak existing ones, or even blend multiple pictures together. It does this all while maintaining consistency in subjects and settings.
It’s the best AI image editor ever created.
How to use it
Head over to OpenArt, pick “Image”, choose “Nano Banana”, and create/edit an image.

Key features
- Strong character consistency across edits.
- Natural language editing — change outfits, background, poses, or even blur or remove elements with ease.
- Supports multi-step edits, blending, and one-shot transformations.
- It can fuse multiple images into a single cohesive scene.
- Works via the Gemini app, API, and even third-party tools like OpenArt.
Best use
Use this when you want to keep people, pets, or objects looking the same while making precise edits.
This model is best for storytelling, visual continuity, and content that shows character evolution across scenes.
My results
I immediately tested Nano Banana as it came out with such hype.
First, I tried to create an image from scratch with it.
Here’s what I told the AI to create: “A portrait of a smiling woman in a jeans jacket against a neutral background.”
Here’s the result:

Then I tried the editing options with this input: “Change the jeans jacket to a floral dress, and place her in a sunny park scene.”
Here’s the result:

That’s so cool! The character looks to me exactly like in the first image.
But let’s not leave it at that. I’ve heard this AI model is good for editing images, so let’s try it on my own photos.
Here’s a selfie of me, a real one:

I’ll now try to change the color of my shirt to red with Nano Banana AI. I just told the AI to “turn the man’s shirt red”.
Here’s what it gave me:

That is crazy! It looks exactly like me, but now the shirt is red.
I mean, I think I don’t need to take professional headshots anymore since AI can do it in 15 seconds…
In fact, I had to try that, so here’s the result after I told the AI to “Make the man wear a dark blue suit and make the background an office.”

Isn’t that impressive?
And by the way, this is not just for editing humans or outfits. You can remove or replace objects in any picture.

I have to admit I couldn’t have believed AI to be this good while looking at the future in 2022 or 2023.
Crazy times!