I've been testing AI image generators for years - and my new favorite surprised me
Why you can trust ZDNET
'ZDNET Recommends': What exactly does it mean?
ZDNET's recommendations are based on many hours of testing, research, and comparison shopping. We gather data from the best available sources, including vendor and retailer listings as well as other relevant and independent reviews sites. And we pore over customer reviews to find out what matters to real people who already own and use the products and services we’re assessing.
When you click through from our site to a retailer and buy a product or service, we may earn affiliate commissions. This helps support our work, but does not affect what we cover or how, and it does not affect the price you pay. Neither ZDNET nor the author are compensated for these independent reviews. Indeed, we follow strict guidelines that ensure our editorial content is never influenced by advertisers.
ZDNET's editorial team writes on behalf of you, our reader. Our goal is to deliver the most accurate information and the most knowledgeable advice possible in order to help you make smarter buying decisions on tech gear and a wide array of products and services. Our editors thoroughly review and fact-check every article to ensure that our content meets the highest standards. If we have made an error or published misleading information, we will correct or clarify the article. If you see inaccuracies in our content, please report the mistake via this form.
AI chatbots like ChatGPT have taken the world by storm because they can generate nearly any kind of text -- including essays, reports, and code -- in seconds. AI image generators, however, are even more fun because they can take a simple prompt and generate a visual representation of whatever you're imagining.
Use cases for text-to-image AI generators can range from personal projects, such as creating greeting cards, event invites, and wallpapers, to professional projects, such as developing brand assets, social media content, or marketing campaigns. The possibilities are endless.
OpenAI's DALL-E 2 made a huge splash as the first mainstream AI image generator. Other companies have released models that surpass DALL-E 2, and OpenAI has since released a more advanced model, DALL-E 3, which replaced its predecessor.
Also: The best AI chatbots: ChatGPT, Copilot and worthy alternatives
I've been testing AI image generators since 2022. For my most recent round of testing, I gave each image generator tool the same prompt: "A vibrant, realistic hummingbird perched on a tree." I also included screenshots to help you decide which tool best matches your needs.
I ranked all the image generators from best to lowest image quality. Since the actual image output is only one factor in choosing an image generator, I also included details on cost, speed, accessibility, and other standouts from my testing.
For image generators whose strong suit is something other than generating high-quality images, such as generating accurate text, I included those results instead as they help you visualize how good it is at that specific task.
What is the best AI image generator overall?
In my most recent test, Google's ImageFX dethroned Microsoft Designer's Image Generator as the best overall AI image generator because it generates the highest-quality, most realistic renditions for free. Google has been a dark horse in the AI space, so the company beating more well-established contenders surprised me.
ImageFX combines accuracy, speed, and cost-effectiveness and can generate images in seconds. A major plus is that the generator is easy to use for beginners and has unique features like expressive chips.
Also: Microsoft launches free mobile app versions of Designer, its AI photo creator
ChatGPT is a solid alternative, especially since OpenAI in August made image generation on ChatGPT available to all free users. ChatGPT is not in first place because you're limited to two photos per day with its image generation tool, which is not a lot compared to most other generators, and because ImageFX produces higher-quality images.
Whether you want to generate images of animals, objects, or even abstract concepts, ImageFX can produce accurate depictions that will meet your expectations. It is highly efficient, user-friendly, and cost-effective.
For the full roundup of the best AI image generators, keep reading.
The best AI art generators of 2024
Google was an underdog in the image generator space, releasing its own AI generator, ImageFX, months after its competitors. The wait was worth it. ImageFX, powered by Imagen 3, can produce high-quality, realistic outputs, even of objects that are difficult to render, such as hands.
ImageFX is a standalone experience, and it's easy to use. All you have to do is sign in to your Google account, type in a prompt, and let it do the magic for you. You can also take advantage of cool features such as "expressive chips," which allow you to swap out elements of your prompts for more generations.
Also: How to use ImageFX, Google's AI image generator
Google also updated its AI chatbot Gemini to let you generate photos using Imagen 3. Like when using ChatGPT and Copilot, you can access the text-to-image model while chatting with Gemini. It's useful to not have to context switch between platforms for your text and image generation needs.
If you are a Google user, you will probably prefer Gemini's user interface because it has the same look and feel as Google Search. This isn't just surface-level. Gemini double-checks the validity of its responses against Google Search and offers links that take you to the search engine.
Imagen 3 in ImageFX: Powered by: Imagen 3 | Access via: Website | Output: Four images per prompt | Price: free
I often play around with AI image generators because they make it fun and easy to create digital artwork. Despite my experiences with different AI generators, nothing could have prepared me for Midjourney.
The output of the images was so crystal clear that I had a hard time believing they weren't photos that someone had taken -- the software has even produced award-winning art.
When it first launched, Midjourney's biggest flaw was that it wasn't very user-friendly. Since then, however, Midjourney has greatly simplified the process, launching a standalone webpage that makes it easy to get started. For step-by-step instructions, check out ZDNET's guide, How to use Midjourney to generate amazing images and art.
The biggest problem with Midjourney is that you must subscribe for $10 per month. Although the rendition of the hummingbird may not seem particularly noteworthy, Midjourney consistently produces high-quality images of everything from objects to people to scenery, which may make the cost worth it to daily users.
Midjourney features: Powered by: Midjourney | Access via: Standalone site or Discord | Output: Four images per prompt | Price: Starts at $10/month
Adobe has been a leader in developing tools for creative and working professionals for decades. As a result, it's not surprising that its image generator is impressive.
Accessing the generator is easy: Visit the website and type a prompt for the image you'd like generated. As you can see above, the rendered image of the hummingbird has impressive detail and is so high-quality it looks like a photo.
Also: This new AI tool from Adobe makes generating the images you need even simpler
This chatbot's biggest standout features are Structure Reference and Style Reference. Structure Reference enables you to input an image that you want the AI model to use as a template. The model then uses this structure to create a new image with the same layout and composition. Style Reference uses an image as a reference to generate a new image in the same style.
These features are useful if you have an image you'd like the new, generated image to resemble, such as a quick sketch you drew or a business logo or style you'd like to keep consistent. When creating a new image, you can also tweak the "strength" of the structure reference (how much the model adheres to the reference's image structure), allowing you to rely on your reference as lightly or heavily as you need.
Another perk is that Adobe Firefly was trained on Adobe Stock images, openly licensed content, and public domain content, making all the generated images safe for commercial use.
Adobe Firefly features: Powered by: Firefly Image 2 or Firefly Image 3 | Access via: Website | Output: Four images per prompt | Price: Free
Microsoft Designer's Image Creator is powered by DALL-E 3, OpenAI's most advanced image-generating model. It produces the same quality results as DALL-E in ChatGPT, but it's free, helping you circumvent the $20-per-month ChatGPT Plus subscription to use DALL-E 3 as much as you'd like.
To access the image generator, visit the Image Creator website and sign in with a Microsoft account, or create one for free if you don't have one.
Another major perk about this AI generator is that you can access it in the same place you access Microsoft's AI chatbot, Copilot (formerly Bing Chat).
This means that in addition to visiting Image Creator on its standalone site, you can ask it to generate images for you in Copilot. To render an image, all you have to do is conversationally ask Copilot to draw you any image you'd like.
Also: How to use Image Creator from Microsoft Designer (formerly Bing Image Creator)
This is convenient: You can satisfy all your image-generating and AI-chatting needs in the same place for free. This facilitates tasks that could benefit from both image and text generation, such as party planning, as you can ask the chatbot to generate themes for your party and then create images that follow the theme.
To do this, visit Copilot's website and log in to access image generation. Note that if you use a school or work account, there may be some limitations that prevent you from generating images, so use a personal account.
Image Creator from Microsoft Designer features: Powered by: DALL-E 3 | Access via: Copilot, browser, and mobile | Output: Four images per prompt | Price: Free
OpenAI, the AI company behind ChatGPT, launched DALL-E 2 in November 2022. The tool quickly became the most popular AI image generator on the market. After launching its most advanced image generator, DALL-E 3, OpenAI discontinued DALL-E 2.
DALL-E 3 is even more capable than the original model, but this comes at a cost. To access DALL-E 3 with no limits, you must be a ChatGPT Plus subscriber, and the membership costs $20 per month. Once subscribed, you can access DALL-E 3 via ChatGPT on the web or the ChatGPT app.
If you are a free user, you can access DALL-E 3 in ChatGPT with a limit of two generations per day, significantly less than other generators on the market.
Also: How to subscribe to ChatGPT Plus (and 5 reasons why you should)
Using DALL-E 3 in ChatGPT is intuitive. Type in whatever prompt you'd like, specifying as much detail as necessary to bring your vision to life, and DALL-E 3 will generate an image that matches your prompt.
The image on the left is the original rendition, and the one on the right is where I prompted it to be "more realistic." Even on the right, it is not as realistic as other models, such as Google's ImageFX or Midjourney.
As you can see in the image at the top of the article, the renditions are the same quality as Microsoft Image Generator because they both use DALL-E 3. Microsoft's tool is free, however, giving it the clear advantage.
Like with Copilot, you can chat and render your images on the same platform, which is convenient for projects that depend on image and text generation.
DALL-E 3 features: Powered by: DALL-E 3 by OpenAI | Access via: ChatGPT website and app | Output: One image per prompt | Price: Limited access for free or a ChatGPT Plus for $20 per month
As seen in the different examples above, many AI image generators can produce incredibly realistic-looking images and, as a result, risk spreading misinformation. As a result, most mainstream image generators have implemented guardrails that prevent users from creating images that replicate public figures or trademarked characters. Leonardo.AI, however, is much more lenient.
Also: The best AI search engines: Google, Perplexity, and more
When I rendered the hummingbird image, the result was comparable to that of DALL-E 3. It didn't stand out among the other options, which is why I didn't include that image.
Leonardo.AI's competitive edge is that you can generate images of characters. We don't recommend that you use an AI image generator to create harmful images of public figures, but there are some benign use cases.
For example, if your friend or kid is a big Disney fan, you can ask the image generator to create a realistic image of Mickey Mouse in a specific environment. The prompt's results are below: "Mickey Mouse on a plane, smiling."
It is also nice to be able to generate images without the safety guardrails that can sometimes be limiting for no reason. For example, if you ask other generators with strict guardrails to generate a person with descriptions similar to a celebrity's appearance, sometimes you won't get a result.
Leonardo.AI features: Powered by: Leonardo's suite of fine-tuned AI models | Access via: website | Output: One to four images per prompt | Price: Free daily allocated tokens, but you must subscribe for more
One of AI image generators' biggest challenges is generating text onto images. Often, the words are misspelled, scrambled, illegible, or not what the user intended. However, Recraft V3 can generate accurate long text strings quickly and for free.
This is particularly useful because, many times, you are using an AI-generated image for a larger project, such as a greeting card or social media post, which could benefit from having text added to it. By being able to generate it when you ask for your initial prompt, you save yourself some time uploading the image into another editor after adding the text you'd like.
Also: I tested this viral AI image generator, and it does text well - finally! Try it for free
Another plus is that when using the image generator, it can make the text appear native to the pic, as seen in the example below where I asked it to generate an image of a computer screen that read ZDNET's brand's mission statement, "ZDNET, tomorrow belongs to those who embrace it today," in electric yellow.
All you have to do is visit the website, click on "Generate AI image," and create a Recraft account or sign in with an existing Google, Discord, Apple, or single sign-on.
Despite originally being named DALL-E mini, this AI image generator is not affiliated with OpenAI or DALL-E. Rather, it is an open-source alternative. Nevertheless, the name somewhat fits as the tool does everything DALL-E does, but with less precise renditions.
Also: How to use Craiyon AI (formerly known as DALL-E mini)
Unlike DALL-E, the outputs from Craiyon aren't as high quality and take longer to render -- approximately a minute, which, all things considered, is still pretty quick. A banner at the top of the page also lets you know whether you should expect delays due to high traffic.
A major plus is that you can continue to tweak the prompt until you get exactly what you envisioned. You can also include "negative words," which help you describe what you don't want to see, ensuring you get the best result. The tool generates nine images, more than any other chatbot listed, albeit at a lower quality.
Craiyon features: Powered by: Craiyon's AI model | Access via: Craiyon website | Output: nine images per prompt | Price: Free, unlimited prompts
Stability AI created Stable Diffusion, the massively popular, open-source text-to-image generator. Users can download the tool and use it for free, but it may require some technical skill.
Also: How to use Stable Diffusion AI to create amazing images
To make the technology more accessible to everyone (regardless of skill level), Stability AI created DreamStudio, which incorporates Stable Diffusion in a UI that is easy to understand and use.
One of the platform's standout features is that it includes many different entries for customization, including a "negative prompt" where you can delineate the specifics of what you'd like to avoid in the final image. You can also easily change the image ratio -- that's a key feature, as most AI image generators automatically deliver 1:1 images.
To access it, you need to create a Stabillity.ai account, or sign in with your existing Google or Discord accounts.
DreamStudio features: Powered by: SDXL 1.0 by Stability AI | Access via: Website | Output: One image per two credits | Price: $1 per 100 credits | Credits: 25 free credits when you open an account; buy more once you run out
This app took the first-place spot for the best overall app in Google Play's 2022 awards, and it has 4.8 stars on Apple's App Store with 142.9K ratings. Dream lets you create art and images with the simple input of a quick prompt.
An added plus is that this AI image generator lets you pick different design styles, such as realistic, expressionist, comic, abstract, fanatical, ink, and more. This helps take the guesswork out of crafting the perfect prompt to get your desired output.
Also: How to use Dream by WOMBO to generate artwork in any style
In addition to the app, the tool has a free desktop mobile version that is simple to use. If you want to take your use of the app to the next level, you can pay $90 per year or $10 per month.
Dream by WOMBO features: Powered by: Wombo's machine-learning algorithm | Access via: Mobile and desktop versions | Output: One image with a free version, four with a paid plan | Price: Free limited access
One of the biggest issues with AI image generators is that they typically train their generators on content scraped from the internet, which means the generators use aspects of creators' art without compensation. This approach also puts businesses that use generators at risk of copyright infringement.
Generative AI by Getty Images tackles that issue by generating images with content solely from Getty Images' vast creative library, with full indemnification for commercial use. The generated images have Getty Images' standard royalty-free license, assuring you that the content is fair to use without fearing legal repercussions.
Another pro is that contributors whose content was used to train the models are supposed to be compensated for their inclusion in the training set. This is a great solution for businesses that want stock photos that match their creative vision but do not want to deal with copyright-related issues.
Generative AI by Getty Images features: Powered by: Nvidia Picasso | Access via: Website | Output: Four images per prompt | Price: Paid (price undisclosed, have to contact the team)
What are AI image generators capable of?
An AI image generator is software that uses a generative AI model to create images from user text inputs, usually within seconds. The pictures vary in style depending on the software's capabilities, but they can typically be rendered in any style you want: 3D, 2D, cinematic, modern, Renaissance, and more.
If you've ever searched Google for hours to find an image you needed, artificial intelligence (AI) may be able to help.
With AI image generators, you can type in a prompt that is as detailed or vague as you'd like. These tools can help with branding, social media content creation, and making invitations, flyers, business cards, and more.
Also: ChatGPT no longer requires a login, but you might want one anyway. Here's why
Another major perk of these generators is that they take seconds to generate, with the longest generation taking about a minute.
Even if you have no professional use for AI, don't worry -- the process is so fun that anyone can (and should) try it out.
Do you need to pay for AI image creators or are the free versions good enough?
Ultimately, whether you should pay for an image generator or not depends on your use cases. If you are a business and need a commercially safe generator or the highest quality renditions, then paying for services like Midjourney or Generative AI by Getty Images makes sense.
If you are a consumer looking to generate images occasionally for fun and leisure, then there are plenty of highly competent free AI image creators, and I would advise you not to pay for a subscription.
How do you use AI image generators?
Like other generative AI models, AI image generators require a lot of data. Typically, these models are trained on billions of images, which are then used to in turn create impressive images.
Follow these simple steps to get started using one yourself:
1) Visit the website of the generator you are interested in using.
2) Sign in or create an account if the site requires you to have one to access the generator.
3) Type a detailed prompt about what you'd like the generator to output.
For a more thorough breakdown on how to start using AI Image generators, check out: How to use Microsoft Designer's Image Creator (formerly Bing Image Creator).
Top AI image generators compared
Google's ImageFX is the best overall AI image generator. Like DALL-E 3, Image Creator combines accuracy, speed, and cost-effectiveness and can generate high-quality images in seconds. Unlike DALL-E 3, ImageFX is entirely free.
Whether you want to generate images of animals, objects, or even abstract concepts, Google's ImageFX can produce accurate depictions that meet your expectations in a user-friendly interface.
AI image generator | Price | Output speed | Accessibility |
Imagen 3 in ImageFX | Free | Fast | Web and app |
Midjourney | Starts at $10/month | Fast | Web |
Adobe Firefly | Free + credits | Fast | Web |
Microsoft Designer's Image Creator | Free | Fast | Web and app |
DALL-E 3 in ChatGPT | Free + subscription | Fast | Web and app |
Leonardo AI | Free + subscription | Fast | Web |
Recraft V3 | Free + subscription | Fast | Web and app |
Craiyon | Free | Slower | Web |
Stability AI's DreamStudio | Free + credits | Fast | Web |
Wombo's Dream | Free + Subscription | Fast | Mobile and Desktop |
Generative AI by Getty Images | Cost determined by chatting with the sales team | Unknown | Web |
Note: Prices and features are subject to change. Since Generative AI by Getty Images is a business tool, I have not tested it, and therefore, couldn't gauge its speed.
Which is the right AI image generator for you?
Although I crowned Google's ImageFX as the best AI image generator overall, other AI image generators perform better for specific tasks. For example, suppose you are a professional using AI image generation for your business. In that case, you may need a tool like Generative AI by Getty Images, which renders images that are safe for commercial use.
On the other hand, if you want to play with AI art generation for entertainment purposes, Craiyon might be the best option because it's free, unlimited, and easy to use.
Choose this AI image generator... | If you want... |
Imagen 3 in ImageFX | The best AI image generator overall, as it creates the highest quality images out of all the free image generators. |
Midjourney | The best AI image generator for high-quality renderings and crystal-clear images. |
Adobe Firefly | The best AI image generator if you have a reference photo you'd like for the AI image generator to use as inspiration in either structure or style when rendering a new image. |
Microsoft Designer's Image Creator | The best AI image generator to access DALLE-3 for free. In Copilot, you can meet all your image-generating needs while chatting with the bot and getting all your questions answered. The image generator is reliable, quick, and accessible. |
DALL-E 3 in ChatGPT | An upgraded version of the original best AI image generator that combines accuracy, speed, and cost-effectiveness. It lets you generate high-quality images quickly and easily, making it an ideal tool for artists, designers, and anyone looking to create unique and original content. |
Leonardo AI | Lenient guardrails that let you generate images of characters. This can be useful for party invites, birthday cards, and more. |
Recraft V3 | The best AI image generator for generating text onto images, including long strings of words. |
Craiyon | The best no-frills AI image generator with unlimited prompts and a straightforward interface. |
Stability AI's DreamStudio | The best customizable AI image generator with a UI that makes it easy to get the exact rendition you want. The fields include size, style, negative prompts, and image prompts. |
Wombo's Dream | The best AI image generator for your phone with multiple templates, realistic renditions, and a mobile app. It also has a free limited access version, making it a great option if you don't want to spend too much money. |
Generative AI by Getty Images | The best AI image generator for business applications. It generates images with content from Getty Images's vast creative library, meaning the images rendered are commercially safe. |
How did I compare these AI image generators?
To find the best AI image generators, I tested each generator listed and compared their performance across UI/UX, image results, cost, speed, and availability. Each AI image generator had different strengths and weaknesses.
UI/UX: The interface design and overall user experience are a big factor. Even though using an image generator should be as simple as typing in a text prompt, some platforms overcomplicate things, making it hard to get started.
Image results: If you use an image generator, it is likely because you want to produce high-quality art, pictures, or other media. Therefore, one of the biggest factors is the ability to create realistic and detailed images with as few artifacts as possible.
Price: There are so many great free AI image generators. We thus want to ensure that if you have to pay to use a model, it is worth every penny, since some free generators can produce incredible outputs.
Speed: Many image generators can produce images in seconds, setting the expectation for image generators to be quick. Therefore, it is noticeable when a model takes longer. Slowness is a significant con.
What's the future for AI image generators?
Many AI image generators on the market shine in terms of speed, quality, and affordability. That said, there isn't much more I'd like to see from image generators that could significantly improve the offerings. Furthermore, the future of AI image generators isn't about updating current offerings, but rather moving from one-dimensional to three-dimensional renditions.
Some companies, such as Stability AI, already have 3D object image generators, such as Stable Video 3D, which generates 3D objects from single images.
Text-to-image video generation is also the next big thing, with many companies, including OpenAI, Google, and Stability AI, already working on their own generators.
Frequently Asked Questions
What is the best free image generator?
Google's ImageFX is the best AI image generator because it produces the highest-quality, most realistic image for free. The photo quality even exceeds that of OpenAI's latest model, DALL-E 3. When accessing DALL-E 3 from OpenAI's services, you need to pay for unlimited access. Otherwise, you're limited to two generations per day.
How do you fix a poorly generated image?
You can fix a poorly generated image by readjusting your prompt to fix the element of the image you are having trouble with. For example, if you say, "Generate an image of the beach," and are disappointed at the way the ocean looks, you can input another prompt that says, "Generate an image of a beach with crystal blue, shallow water." The more specific you get, the better your result will be.
Are there ethical implications with AI image generators?
AI image generators are trained on billions of images from across the internet. These images are often artworks belonging to specific artists, which an AI tool repurposes to generate your image. Although the output is different, the new image has elements of the artist's original work that are not credited to them.
Are AI-generated images copyrighted?
Whether an AI-generated image is copyrighted depends on the terms of the generator you are using. The copyright policies also vary between subscription tiers. The best way to proceed is not to assume and to double-check before you use or publish a generated image.
How do I properly disclose that my images were AI-generated?
A good habit when using AI to generate images is to disclose that AI was involved in the process. Doing so helps build trust with your audience, as well as helps prevent misinformation from spreading. A disclosure can be as simple as something that reads, "Generated by [the name of the generator]."
You will notice that at ZDNET, we always disclose when we use an AI image generator image in an article, using a disclosure such as, "Sabrina Ortiz/ZDNET via [Image generator name]."
Are there DALL-E 3 alternatives worth considering?
Many AI image generators produce better results than OpenAI's tool. If you want to try something different, check out one of our alternatives above or the two additional options below.