Google's new AI image model has earned the nickname “Nano Banana” after a series of fruit-themed teasers from executives and early reviewers considered its ability to be bananas. Since its release, the nickname “Nano Banana” has been stuck. But don't worry if you don't know exactly what it refers to. Here's what you need to know to find a model and get the most out of this powerful new tool:
The model's government name is a Gemini 2.5 flash image. It doesn't sound as fun as “nano bananas,” but it's important to be aware of it. He has left dramatically over the past few weeks, especially compared to other AI imaging tools. Google reported that users have created over 5 billion AI images, and that the model led to over 10 million new Gemini users in the first few weeks after the release of Nano Banana. It also helped me launch Gemini at the top of the Apple App Store's free app chart for a while.
While Google has invested in AI integration everywhere, Nano Banana Image Editing appears to be one of Google's biggest leap forwards to develop the best AI.
As an expert in AI image generators at CNET, I put Nano Banana in my test to see if the model holds on or if everything is simply an AI hype. Google's Nano Banana is impressive and has a refreshing focus on editing existing photos. There were still some slip-ups, but this is expected with AI services. However, adding new elements to existing photos was painful and good. If you're an AI enthusiast, or even those who need quick custom photo editing, I think it's worth exploring.
To save time, this guide is built on my real-life experience finding and playing around with models, as well as crowdsourced tips to get the best results. For more information, see our guide to creating AI images and creating the best AI image prompts.
How to access Google's Nano Banana
There are several ways to access nanobananas. If you want to go directly with Google, you will need Google's AI Pro plan, an AI-specific subscription that costs $20 per month. Once you've signed up, you can go to Google's AI Studio and choose a 2.5 Flash image as the model you want to use. Unlike Openai's image generation using ChatGpt, you cannot edit photos through Gemini chat.
Developers can also access the models via Vertex AI and Gemini APIs. If you don't know what they are, you probably won't be able to access them. So don't worry about that. Stick to an AI studio that is open to everyone.
You can also use Nano Banana using certain Adobe programs. Nano Banana is available as a third-party model on Photoshop, Adobe Express and Firefly, an AI platform. These programs are also usually paidwalled, but if you already have some Adobe programs, it's worth checking if you have access to them. As many third-party models are available on Adobe,[生成]Please select 2.5 flash images before clicking.
As always, an important privacy note: Google's general Gemini privacy policy states that information you upload can be used to improve your AI products. Therefore, the company recommends avoiding uploading sensitive or personal information. Also, please make sure you use AI when sharing those images, as photos from the nanobanana editing are not watermarks.
Don't miss out on unbiased technical content and lab-based reviews. Add CNET As a preferred Google source.
How to use Google's Nano Banana and make the most of it
If you're at Google's AI Studio, you can easily upload images for editing. But like any AI imaging tool, the chances of success depend on how well you are driving your model. Here are some tips for using Nano Banana.
- specifically. Creating an AI prompt is different from Google search. The more you provide the details, the more likely you will get the results you need. We'll talk about colour, style and aesthetics. The keywords are great, but don't worry about writing.
- Tell it what you don't want to change. This is different to the advice I give when working with AI image generators. Most AIS have a hard time figuring out negative prompts or understanding what you don't want to see. However, nano bananas are good for understanding what you're doing and don't want to adjust them, so include it in the prompt.
- If you need text, please tell us exactly what you want. This is partly inspired by Google's own best practices for prompts, but it is also something I saw in my own tests. You need to be specific when you need the text. Otherwise, it will be messy. Don't forget to include the font style and color as well.
Here is an example of how to edit an existing photo (left) in Nano Banana to include new elements.
One of the best things you can use with Nano Banana is to add new elements to the images you already have. However, there are a few other good use cases. Based on some of my favorite prompts I use to test AI services, here are some ideas that can help you inspire.
- Play with variations. You can ask the model to create different versions of existing photos. It's a headshot of a different hairstyle or change the photo to a different style. For example, I took photographs of natural landscapes and rethinked AI in the style of dark, sci-fi and fantasy worlds.
- Use Nano Banana as an AI-equipped eraser. Removing background distractions is one of the most common uses of photo editing software. A weight-worthy AI should help remove false obstacles and seamlessly fuse the affected area with the photograph.
- When you're stuck, ask Gemini for inspiration or ideas. I always recommend talking to real people through photo editing projects when you can. But if you can't, you can use the chatbot as a kind of soundboard for your ideas. Gemini can also help you put your own ideas into the words that nano bananas can understand.
For more information, check out Google's Nano Banana and the hands-on experience with the best AI image generators.
