Table of Contents
Generative artificial intelligence has made it easier than ever to create content. These AI image services let you enter a simple text description of the image you want (called a prompt) and it almost instantly spits out a series of images that reflect your artistic vision. AI creative services aren’t holistic enough to replace professional human creators, but they can be used to gain inspiration, mock up ideas, and create personalized art. There are many AI image generators on the market, but not all of them will produce good results.
The AI image generators on this list have been thoroughly tested. CNET reviewers spent months with these programs, generating hundreds of images and creating everything from cartoon rainbow safaris to dramatic science fiction scenes and photorealistic stock photos. At some point, every service on this list spits out a shaky or unusable image. The test of a truly superior AI image generator is how well equipped it is to handle these quirks and fix flaws. Editing tools and customization options are a big part of that, which is why we test them extensively. Privacy policies, including how generators may train on your content, are also important considerations when selecting the best services. This is what made the cut.
Read more: AI Image Creation: A Complete Guide with Expert Advice
What is the best AI image generator?
Dall-E 3 from OpenAI is our top pick for AI image generators. It can handle long, complex questions, gives you plenty of editing and customization options, and its unique conversation flow makes it easy to bring your artistic visions to life. If you don’t want to sign up for the $20 per month ChatGPT Plus plan, Leonardo AI offers great AI visuals in its free plan. Beginners will find Canva’s user-friendly design easy and accessible to use, and professional creatives looking to experiment with AI art should try Adobe Firefly and its family of AI tools.
Best AI image generator of 2024
Like
- Understands long, complex questions
- Captivating, dynamic images
- The conversation style allows for easy adjustments
Don’t like
- Photorealistic images can appear fake
- Slow to generate images
OpenAI’s Dall-E 3 is CNET’s 2024 Editor’s Choice for Best AI Image Generator. Dall-E 3 is great for amateurs and professionals, with plenty of editing tools such as different styles and adjustments that you can take advantage of. Dall-E is unique among its competitors because of its conversational flow, which makes it easy to send follow-up requests, and its ability to handle complex queries. OpenAI may use your content to improve its models, but the company says it won’t share your data with advertisers and offers a way to request to opt out of model training. With the $20 per month ChatGPT Plus subscription, Dall-E 3 creates vibrant, engaging visuals with limited AI quirks.
Read our Dall-E 3 review.
Like
- Clear and detailed images
- Useful, fast technical tools
Don’t like
- No post-generation editing tools in the free plan
- Mediocre privacy policy
Leonardo is great for AI creatives on a budget. It has an extensive free plan that will give you enough generation credits quickly. It doesn’t have any free post-generation editing tools (they have a paywall), but it does have a prompt improvement tool and other customization elements that can guide you in building the best prompts. Leonardo makes it easy to discover what’s possible with AI and create vivid, detailed AI images. The free plan easily knocks its competitors out of the water, even the easy-to-use Canva.
Read our Leonardo AI review.
Like
- Excellent artistic styles
- Good refinement options
- Competitive generation speed
Don’t like
- Struggles with photorealistic images
- Difficulty when elements contact or interact
- Does not accept complex questions
Adobe Firefly’s generative AI imaging tools are built directly into Adobe Creative Cloud, including Photoshop, making it a great option for professional creatives who want to experiment. Firefly offers many stylistic and artistic options, and its refinement tools are similar to editing software that creatives will be familiar with. Firefly is trained on Adobe’s proprietary Stock Catalog, which contains high-quality licensed public domain content. Firefly does not train on your content. If you already pay for an Adobe Creative Cloud subscription, Firefly can be an easy way to develop ideas or get inspiration.
Read our Adobe Firefly review.
Like
- Excels at creating fun, quirky and cartoon images
- Easy to use on desktop and in the mobile app
- Does not share your content to train its AI
Don’t like
- The free plan has a hard limit on the number of images you can create
- Difficult to generate non-square images
Canva is an extremely easy-to-use graphic design service, and the 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 it’s great for those on a budget and time constraints. Canva’s privacy policy is particularly secure because Canva doesn’t 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 in the mobile app. It’s a simple, easy-to-navigate AI image generator, perfect for beginners and Canva enthusiasts.
Read our Canva review.
How we tested AI image generators
CNET takes a hands-on approach to reviewing AI image generators. Our goal is to determine how good it is compared to the competition and what purposes it best serves. To do that, we provide the AI prompts based on real-world examples, such as rendering in a certain style, combining elements into one image, and processing longer descriptions. We rate the image generators on a 10-point scale, taking into account factors such as how well images match clues, creativity of the results, and response speed. To see how we test AI for more.
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 question and how clearly elements and details are displayed. For example, a service that closely matches the sci-fi/fantasy spaceship scene described in a prompt is considered accurate, as is a service that generates an image of a human with a distinct expression.
On the other hand, hallucinationsor the way AI makes things up are also important to consider. Hallucinations for AI images often appear in the strange quirks or flaws you didn’t ask for, such as people with three arms or disappearing elements. creativity is an inherent necessity of an AI image generator, but hallucinations are obvious blunders and not whimsical or unrealistic style elements. All AI image generators encounter these problems, but great AI services have them less often and offer editing tools ready to solve them.
The number clarifying instructions required indicates how much work you need to do to get the image you want. If you can’t follow through on an operation or additional request, that could be a red flag or annoyance that you should be aware of. Generators that closely follow directions and provide editing tools make it easier to bring your vision to life.
Reaction speed is also important to consider. Most image generators are quite fast, less than two minutes. Services that can generate images faster, between 10 and 30 seconds, give an edge to their user experience.
Other AI image generators we tested
Midjourney is a solid option for an AI image generator, but it didn’t make our top pick because it’s currently only available on Discord, is paid only, and inconsistently matches directions. Midjourney offers nice upscaling or editing tools for individual images, but you’ll have to use them often. It’s also worth noting that all your images are public and accessible in an online gallery unless you take them in stealth mode, which is only available on the more expensive Pro and Mega plans.
Google’s ImageFX also didn’t make our top pick due to persistent inaccuracies and “overbearing babysitting” that results in harmless prompts being rejected. All AI generators struggle with photorealistic faces and hands, but ImageFX also struggles with inanimate objects and logos. This wouldn’t be a dealbreaker, except that the editing and refinement tools aren’t up to par for making images usable.
Check this out: How to use ChatGPT’s new voice conversations
OpenAI’s Dall-E 3 is CNET’s Editor’s Choice for the Best AI Image Generator for its extensive editing and adjustment tools, close guidance following, and level of detail in images. Dall-E uses a conversational flow and handles more complex instructions very well.
There are a number of free and freemium AI image generators to choose from, but we recommend starting with Leonardo AI or Canva. Both services offer free AI image generation with fast response speeds and appropriate, fast compliance.
To create the best AI image, you need to build the right image prompt. As you write what you want your generator to create, make sure to include the style, aesthetic, dimension, and any specific elements you want to include. Some generators have trouble following specific prompts, but it’s better to start with a specific prompt and refine or scale back as necessary.
There are many reasons why your AI image generator may not give you what you want. Try customizing your prompt, using customization settings and editing tools to eliminate any quirks or conflicting elements. Unfortunately, sometimes the only option is to scrap your idea and start over.
Yes. If you use an image created entirely by an AI image generator, you must disclose this wherever you share it.
