Innovations in generative AI mean you no longer need to be a Photoshop expert to create custom content. AI image generators like Google’s nano banana, Midjourney and Stable Diffusion can bring your artistic visions to life in minutes using a simple text description called a prompt. But not every AI image generator is going to be right for your project, budget and skill level. We tested several of them to figure out the right service for you.
Our Picks
PHOTO EDITING SOFTWARE DEALS OF THE WEEK
- $40 (save $25)
- $100 (save $40)
CNET reviewers have extensively tested each service on this list. We’ve generated everything from photorealistic stock images to fantastical sci-fi scenes to cartoon-style graphics. A good AI image generator can handle a variety of styles, but a great AI image generator will also give you the tools to fix errors when they inevitably pop up. Customizable settings or hands-on controls are a great perk for professionals, whereas beginners may be more comfortable using a chat-to-edit workflow. We also closely inspect each service’s privacy and usage policies to understand how a company uses your data.
AI image and video generation is a growing and controversial industry. Creators, artists and copyright owners have voiced serious concerns about how AI models are trained, sometimes on their copyright-protected work, allegedly without their permission and compensation. On the flip side, there are also worries that the guardrails and policies companies have in place aren’t enough to prevent the creation of abusive and illegal AI content. Because of this, it’s important you check a service’s policies before you use it and disclose your use of AI if you post AI images online.
Here are our top AI image generators. For more, you can check out the top AI video generators and our AI essentials.
From left to right, this is how OpenAI, Adobe Firefly Image Model 5 and Gemini 2.5 Flash Image (nano banana) interpret the same prompt of a dark-haired girl sitting at a piano.
Don't miss any of our unbiased tech content and lab-based reviews. Add CNET as a preferred Google source.
What’s the best AI image generator?
The best AI image generator for you is going to depend on your project, budget and skill level. Google’s nano banana pro (formally named Gemini 3 Pro Image) is our best overall choice because of its ability to create realistic images and legible text in images. But if you’re a professional creator, you’re going to want to use Adobe Firefly, which guarantees all your AI content is commercially safe, meaning you don’t have to worry about copyright infringement concerns. Beginners should stick with the ultimate user-friendly option, Canva.
You can see here how each model interpreted the same prompt differently. Nano banana (left) used a more fantasy aesthetic while Firefly 5 (right) has a more photorealistic look.
Pros
- Excellent character consistency and realism
- Can edit existing images as well as generate new ones
- Can create legible text in AI images
Cons
- Longer generation time
- Info in graphics may be inaccurate
Google’s nano banana models took the AI industry by storm in 2025. The original model was praised by fans for its ability to maintain character consistency, and the new pro model is even more capable of handling image editing and generation. Nano banana pro is the best program for generating text in images, like infographics. Even though you’ll want to closely inspect those images to make sure those facts are correct, it’s miles ahead of any other AI image generator.
Nano banana pro is available in Gemini and Google’s other platforms, like Flow and NotebookLM. It’s a great choice for anyone who wants to create an AI image; its level of detail and photorealistic style adherence is unmatched. The results are scarily realistic-looking, so make sure you disclose your AI usage when you use them. You can create a few images for free (select the “Thinking” pro model) before you’ll need to upgrade to a subscription, which begins at $20 per month.
Pros
- Extremely easy to create and edit images
- Able to handle many different aesthetics/styles
Cons
- Free plan has a hard generation limit
Canva is an extremely user-friendly graphic design service and its AI image generator, Magic Media, is a great option for amateur AI creators. Magic Media is a minimalist service, which isn’t great if you need extensive editing tools but is good for folks on a budget and time crunch. Canva’s privacy policy is notably secure, as Canva does not train its AI on your content, and the images you generate are always private, unlike many competitors. Canva also makes it easy to integrate your AI images into your other projects, on desktop and on the mobile app. It’s a no-frills, easy-to-navigate AI image generator perfect for beginners and Canva lovers.
Pros
- Commercially safe outputs
- Integrated with and exportable to other Adobe programs
Cons
- Struggles with photorealistic images
Adobe Firefly’s family of generative AI image tools is built directly into Adobe Creative Cloud, including Photoshop, which makes it a great option for professional creatives looking to experiment. Firefly offers a lot of stylistic and artistic options, and you can take advantage of its robust pre-generation settings to set the aspect ratio and style. Firefly models are trained on Adobe’s own Stock catalog, which includes high-quality licensed and public domain content. Firefly does not train on your content, and its outputs are commercially safe. If you’re already paying for an Adobe Creative Cloud subscription, you might already have access to Firefly. You can also use non-Adobe AI models through Firefly. The Firefly hub can be an easy way to mock up ideas or spark inspiration.
Pros
- Fast generation time
- Stability AI offers extensive editing tools
Cons
- Complicated availability because it’s open source
Stable Diffusion is a household name in AI image generation for a reason. This family of AI image models is open source, and they’re great at creating detailed and colorful AI images. If you use Stability AI, you'll have access to some of the most comprehensive editing tools we've seen in AI image generators. There are a lot of different ways to access Stable Diffusion models -- that's the nature of open source -- including through Stability AI's programs, other third-party platforms and running models locally on your own device. The privacy policy and price will depend on what method you pick. Stability AI offers a $10 per month plan and lets you opt out of model training.
Pros
- Extremely creative results, versatile stylistically
- Community focus with an active Discord server
Cons
- Too realistic generations of protected characters
Midjourney is a solid option for an AI image generator, but there’s one area where it stands above the rest. Midjourney is the most creative option. That makes it a great choice for brainstorming, storyboarding or other kinds of creative work. But if you’re looking for ultrarealistic or stock photography-like imagery, you won’t get what you want. Improvements to Midjourney have helped with its prompt adherence, and it offers nice upscaling or editing tools. You’ll want to note that all your images will be public and accessible in an online gallery unless you create in stealth mode, which is only available in paid plans. Midjourney has been in the news a lot lately, as Disney, Universal and Warner Bros. are suing the company, alleging that its ability to create AI versions of its recognizable characters is copyright infringement.
Pros
- Available with free or paid ChatGPT account
- Very creative images, sometimes has good text in images
Cons
- No advanced editing tools
OpenAI’s AI image model is great for free or paying ChatGPT users. The AI images are creative, engaging and can mimic a variety of artistic styles. The company says it’s great at generating legible text in images, but our testing revealed it can be very hit-or-miss. A lack of post-generation editing features makes it hard to fix those mistakes when they happen. You can turn off AI training and upgrade to its $20 per month plan as needed. OpenAI’s image generation is an affordable option for those working on a budget or ChatGPT users who need artistic assistance.
(Disclosure: Ziff Davis, CNET's parent company, in April filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)
How we tested AI image generators
CNET takes a practical approach to reviewing AI image generators. Our goal is to determine how good it is relative to the competition and which purposes it serves best. To do that, we give the AI prompts based on real-world use cases, such as rendering in a particular style, combining elements into a single image and handling lengthier descriptions. We score the image generators on a 10-point scale that considers factors such as how well images match prompts, creativity of results and response speed. See how we test AI for more.
AI image generators compared
| CNET TOP PICKS | BEST FOR | PRICE | STYLE | PARENT COMPANY | LATEST MODEL |
|---|---|---|---|---|---|
| Nano Banana Pro | Best choice overall for any AI image generation | Free, options to upgrade starting at $20/month | Excels at character consistency, realism | Nano Banana Pro (Gemini 3 Pro Image) | |
| Adobe Firefly | Professional creators | Plans start at $10/month, credits sometimes included in other Adobe plans | Versatile, great for any project | Adobe | Adobe Firefly Image Model 5 |
| OpenAI | ChatGPT users | Free, options to upgrade starting at $20/month | Good at cartoon-ish styles | OpenAI | GPT-Image-1 |
| Canva | Beginner or non-professional creators | Free, options to upgrade starting at $13/month | Can handle a lot of different styles | Canva | Canva AI/Magic Media |
| Stable Diffusion | Open-source AI creation | Depends on method, Stability AI is $10/month | Excels at a variety of styles, including realistic | Stability AI, open source | Stable Diffusion 3.5 |
| Midjourney | Brainstorming and other creative work | Plans begin $10/month | Excels at highly detailed, non-realistic styles | Midjourney | Midjourney V7 |
Factors to consider
Accuracy is an important factor to consider when evaluating AI image generators. Accuracy in this case refers to how well a service matches your prompt and how clearly elements and details are rendered. For example, a service that closely matches the sci-fi/fantasy spaceship scene described in a prompt would be considered accurate, as would a service that generates a picture of a human with a clear expression.
On the flip side, hallucinations, or the way AI makes stuff up, are also important to consider. Hallucinations for AI images tend to be seen in the weird quirks or flaws that you didn’t request, like people with three arms or disappearing elements. Creativity is an inherent necessity of an AI image generator, but hallucinations are obvious flubs, not whimsical or non-realistic style elements. All AI image generators run into these problems, but great AI services have them less frequently and offer editing tools ready to fix them.
The number of clarifying prompts required indicates how much work you’ll have to put into getting the image you want. If you can’t follow up with an edit or additional request, that can be a red flag or annoyance to look out for. Generators that adhere closely to prompts and offer editing tools make it easier to bring your vision to life.
Response speed is also important to consider. Most image generators are pretty quick, less than two minutes or so. Services that can generate images quicker than that, between 10 and 30 seconds, bring an edge to their user experience.
We also pay close attention to an AI company’s policies. Many groups, including celebrities, researchers and copyright owners, are concerned with AI’s ability to create content that could potentially infringe upon others’ legal protections. Our testing of newer AI generators includes seeing how they respond to prompts that ask for specific brand names and products, whether that’s an iPhone or the Teenage Mutant Ninja Turtles. Often these prompts are rejected. Moderation systems and privacy policies are integral parts of your user experience, and we highly recommend you review them before using any AI program.
You can see Midjourney (left) really leans into its signature high fantasy style while Canva (right) goes for an image that’s somewhat more realistic (or as realistic as unicorns can be).
Other AI image generators we tested
A previous version of this article selected OpenAI’s Dall-E 3 as our best overall pick. It was also a CNET editor’s choice award winner in 2024. There is still a lot to like about Dall-E, including its conversational flow that makes edits easy. But explosive growth and innovation in the field of AI imagery throughout 2025 have led to better models. You can still use Dall-E, now up to its fifth generation model, and it’s a good choice for ChatGPT users who need quick, simple AI image generation.
Leonardo.Ai was one of our favorite AI image generations in 2023. It offers helpful prompt engineering tools for beginners, has a generous generation credit policy and makes vivid results. In 2024, Leonardo.Ai was acquired by Canva, and Canva said it would integrate Leonardo’s tech into its in-house models. You can still use Leonardo.Ai through its solo website.
