In this guide, we'll show how to use Nano Banana step by step and explain which Nano Banana version to choose and how much it costs. Let's get started.
In this guide, we'll show how to use Nano Banana step by step and explain which Nano Banana version to choose and how much it costs. Let's get started.

Nano Banana is the nickname of Google's AI image models. Officially, these models are called Gemini Flash Image, but before releasing the first iteration, Google tested it anonymously on LMArena as "nano-banana", and the name stuck.
You can use Nano Banana to create new images from a text description or to edit your own photos. The table below shows all Nano Banana versions Google has released so far:
| Model | API model ID | Released |
|---|---|---|
| Nano Banana 2.1 | gemini-nano-banana-2.1 | October 6, 2026 |
| Nano Banana 2 | gemini-3.1-flash-image | February 26, 2026 |
| Nano Banana 2 Lite | gemini-3.1-flash-lite-image | — |
| Nano Banana Pro | gemini-3-pro-image | November 20, 2025 |
| Nano Banana | gemini-2.5-flash-image | August 26, 2025 |
The current versions are made for different tasks:
You can use Nano Banana in three types of apps:
Below, we show how to use Nano Banana in the Gemini app and in Overchat AI, step by step.
To get started, open a page for image generation in Gemini. Go to gemini.google.com, open the sidebar and click Images.
Here, you can get started by writing a prompt or by using ready-made templates, like Marble and Chibi. If you upload a photo, you can also edit it with Nano Banana.
Let's start by creating an image from text. To do this, click + next to the prompt box and choose Create image.

Now, describe the image you want to create. Write in full sentences, the way you'd describe the picture to a photographer: who is in the frame and what the light is like.
For example, we used this prompt from Google:
A photorealistic close-up portrait of an elderly Japanese ceramicist with deep, sun-etched wrinkles and a warm, knowing smile. He is carefully inspecting a freshly glazed tea bowl. The setting is his rustic, sun-drenched workshop. The scene is illuminated by soft, golden hour light streaming through a window, highlighting the fine texture of the clay. Captured with an 85mm portrait lens, resulting in a soft, blurred background (bokeh). The overall mood is serene and masterful.
Then choose the aspect ratio in the menu under the prompt box and send the prompt. We chose 4:3.
Gemini created the image in about 20 seconds. Warm window light falls on the man's face, and the shelves behind him are blurred, just like we asked in the prompt.

If you want to change something in the image, just write it in the next message. You don't need to repeat the whole prompt. It works best when you change one thing at a time.
For example, we asked Gemini to change the glaze on the tea bowl to deep cobalt blue and keep everything else the same. Only the bowl changed, and the rest of the image stayed as it was.

Now let's edit a real photo. Click + next to the prompt box, choose Upload files and describe what you want to change.
We uploaded a photo of a woman in a denim jacket on a city street and asked Gemini to change the jacket and the season:
Change her denim jacket to a camel wool coat and make it a snowy winter day on the same street. Keep her face, hair and pose exactly the same.
In the edited photo, the woman looks the same, but now she's wearing a camel coat, and there's snow on the street.

When you're happy with the result, click the download icon in the top right corner of the image. On the free plan, you can download images in 1K. On paid Google AI plans, you can download them in 2K.
If you have a paid plan, you can also recreate the image with Nano Banana Pro. To do this, click More under the image and choose Redo with Pro.
Overchat AI is an all-in-one AI app with Nano Banana and other image models, like GPT Image 2.5 and Seedream 5 Pro. Compared to the free Gemini app, it gives you more control over the result:
To get started, open the image generator in Overchat AI and click the model name in the prompt box. In the Models tab, you'll see four Nano Banana versions and their prices in credits:
| Model | Credits per 1K image |
|---|---|
| Nano Banana 2.1 | 100 |
| Nano Banana 2 | 100 |
| Nano Banana Pro | 150 |
| Nano Banana | 15 |
For this guide, we chose Nano Banana 2.1, the newest version.

If you want to compare versions, turn on Multiple selection and tick several models. You'll get one image from each model with the same prompt.
Next, add the photos you want Nano Banana to use. Click Add References and upload them from your device. You can also pick images you created earlier in Overchat AI.
When you write the prompt, refer to each photo by its tag, like @Image1 or @Image2. This way, Nano Banana knows what to take from each photo.
For example, we added the photo of the woman from the Gemini step and a photo of white sneakers. Then we asked Nano Banana 2.1 to combine them into a fashion ad:
Create a fashion ad photo: the woman from @Image1 wearing the white sneakers from @Image2, sitting on the front steps of a brownstone on an autumn morning. Keep her face, hair and smile the same.

Then click the settings icon next to the model name. Here, you can change the image settings:
If you choose a higher resolution, the image costs more credits. For example, a Nano Banana 2.1 image costs 100 credits in 1K and 144 credits in 2K. We chose 2K and a vertical 3:4 frame.

Now click the button with the credit price to generate the image. It will appear in the feed above the prompt box.
Nano Banana 2.1 kept the woman's face and her denim jacket from the first photo and put the sneakers from the second photo on her feet.

If you want to change something, open the image and describe the change in the prompt box under it. You can also use the editing tools at the top, like Erase and Change background.
To animate the image, click Make video. The image becomes the first frame of the video, and you can choose a video model like Kling 4.0 or Seedance 2.5.
In Gemini
On the free plan, you can create a limited number of images a day and download them in 1K. On paid Google AI plans, the daily limit is higher, and you can download images in 2K. Here are the prices of all plans:
| Plan | Price | Image downloads |
|---|---|---|
| Free | $0 | 1K |
| Google AI Plus | $4.99/month | 2K |
| Google AI Pro | $19.99/month | 2K |
| Google AI Ultra | $99.99/month | 2K |
In the Gemini API
If you'd like to generate via the Gemini API, the pricing is per image. Here's a table that breaks down the prices:
| Model | 1K image | 4K image |
|---|---|---|
| Nano Banana 2.1 | $0.0336 | $0.113 |
| Nano Banana 2 | $0.067 | $0.151 |
| Nano Banana Pro | $0.134 | $0.24 |
In Overchat AI
On the free plan, you get 100 credits a day, which is enough for one Nano Banana 2.1 image in 1K. On paid plans, you get 2,000 credits a week or 8,000 credits a month, depending on the plan. Here's a table for more details:
| Plan | Price | Credits |
|---|---|---|
| Free | $0 | 100/day |
| Weekly | $5.99/week | 2,000/week |
| Monthly | $14.99/month | 8,000/month |
| Yearly | $49.99/year | 8,000/month |
On the monthly plan, one Nano Banana 2.1 image costs about $0.19 in 1K and $0.27 in 2K.
These tips work both in Gemini and in Overchat AI:
Yes, you can use Nano Banana for free in the Gemini app with a Google account. On the free plan, there's a daily limit, and you can download images in 1K. In Overchat AI, you get 100 free credits a day, which is enough for one Nano Banana 2.1 image.
No. Nano Banana is an image model, and you can use it inside the Gemini app. You can also use it in other apps, like Overchat AI.
The newest model is Nano Banana 2.1. Google released it on October 6, 2026. It replaces Nano Banana 2 in the Gemini API and costs half as much per image.
Nano Banana 2.1 is the main model for everyday images and edits. Nano Banana Pro is made for professional images with a lot of text, like posters and infographics, and it costs more.
In Overchat AI, you can add up to four reference images to one prompt with Nano Banana 2.1. If you use the Gemini API, you can send up to 14 reference images in one request.
Yes, but you won't see it. Every Nano Banana image has an invisible SynthID watermark that shows it was created with AI. The image we downloaded from Gemini on the free plan had no visible watermark.