Top AI-Powered Image Generators of 2023

Create amazing pictures with AI that turns your words into images!

Glenn Burgess

AI image generators have been making a big splash for over a year now. If you’ve been on social media, watched TV news, or flipped through a magazine, you’ve probably seen AI-created images all over the place.

They’re everywhere, and it’s no surprise why: these tools have gotten really good and are accessible to everyone. If you want to join in the fun or enhance your business with AI-powered tools, the apps on this list have got you covered.

That’s about as long as people have been seriously thinking about these tools, and it’s incredibly exciting to see how much they’ve improved. Instead of diving into complex discussions about art and copyright issues in their training data, let’s focus on the fact that these AI image generators can now create amazing images based on written instructions.

It’s worth spending some time experimenting with one of these text-to-image AI apps, even if it’s just to admire their technical wizardry.

The best AI image generators

Here are some user-friendly AI image generators to consider:

DALL·E 2 – Simple and creative.
Midjourney – For top-notch AI image results.
DreamStudio (Stable Diffusion) – Customize and control your AI images.
Firefly (Photoshop) – Easily integrate AI-generated images into your photos.

How do AI image generators work?

Indeed, these AI image generators can work their magic by taking your text prompts and transforming them into corresponding images. This means you can let your imagination run wild, whether you’re requesting “a Canadian man riding a moose through a maple forest in an impressionist oil painting style,” “a cozy Irish wolfhound savoring a pint in a Vermeer-style pub painting,” or even “a photograph of a donkey exploring the moon.” The possibilities are virtually limitless!

You’re absolutely right! The limits of AI image generators are mainly defined by your creativity, the AI’s ability to understand your prompt, and any safeguards in place to prevent inappropriate content from spreading online.

Most AI image generators follow a similar process. They’re trained using vast datasets of image-text pairs to teach them what different objects and concepts look like. This training helps them interpret a wide range of prompts, although there’s an art to crafting prompts that yield accurate results.

When it comes to creating AI-generated images, the latest models use a technique called diffusion. It’s like starting with a random pattern (imagine a cloudy sky) and then gradually refining it in a series of steps to match the AI’s interpretation of your prompt. Think of it as looking at a cloud that vaguely resembles a dog and magically turning it into a more and more convincing dog shape with each step. This process allows for incredible creative possibilities!

It’s important to set realistic expectations when using text-to-image generators. While they can achieve some impressive results, they might not entirely replace the need for traditional product photoshoots or quickly provide header images for blog posts.

Sourcing images from stock photo websites can be a faster and more reliable option in certain cases. Although these AI models are talented, they might not always match the precision of a custom photo, especially when it comes to intricate details.

What makes the best AI image generator?

The surge in popularity of AI image generators in the last 18 months is quite remarkable, especially considering that not long ago, their results were far from impressive. The underlying technology was impressive from a research perspective, but the images they generated were often lackluster. Even the original DALL·E, when it launched in 2021, was seen more as a fun novelty than a groundbreaking innovation.

Because AI image generators are relatively new, there aren’t many exceptional options to choose from. The four major players on this list—DALL·E 2, Midjourney, Stable Diffusion, and Adobe Firefly—are here for different reasons, but they are also the only apps that met my specific inclusion criteria for two important reasons:

I focused on apps that enable you to generate AI images from a text prompt. While tools that ask you to upload a collection of your photos and then generate AI portraits can be entertaining (often built using Stable Diffusion), they don’t fall into the category of general-purpose image generators that I was considering.
I evaluated the AI image generators themselves, rather than tools built on top of them. For instance, NightCafe is an AI picture generator with a thriving community and app, but it merely facilitates the use of algorithms like DALL·E 2 and Stable Diffusion (along with some older ones). While it’s worth exploring, it doesn’t meet the criteria for this list.

Additionally, I took into account the ease of use, customisation options (such as AI image upscaling), pricing models, and, most importantly, the quality of the results. I have been using and writing about DALL·E 2, Midjourney, and Stable Diffusion since their launches, and I’ve been covering Photoshop for over a decade.

So, I’m quite familiar with how they work, along with their various strengths, weaknesses, and quirks. Comparing all four of them using the same prompts was a fascinating experience, and I’m thrilled to say that each one offers compelling reasons to use them.

One more thing to note before we proceed: all these tools are technically in beta, and it’s likely they will remain in that status for some time. While they continue to improve steadily, AI image generators still have a long way to go before consistently producing outstanding results and seamlessly integrating into commercial workflows.

Please be aware that, unfortunately, being in beta doesn’t mean these tools are free. Since the initial writing of this article, both DALL·E 2 and Midjourney have discontinued their free trials. This change is understandable due to the substantial computing resources required for AI image generation, but it does make it less convenient for casual experimentation.

The best AI image generators at a glance

The best AI image generator for ease of use

DALL·E 2

DALL·E 2 stands out as a prominent player in the realm of AI image generators, and its reputation is well-deserved. It was a trailblazer, being the first AI-powered image generator to produce captivating and attention-grabbing images that went viral. While being the first doesn’t always make something the best, DALL·E 2 has a lot to offer.

One of its standout features is its remarkable user-friendliness. It’s incredibly straightforward to use—simply type in your vision, hit the Generate button, and within moments, you’ll have four AI-generated variations to select from.

Though DALL·E 2 no longer provides a free trial, it remains a worthwhile choice at $15 for 115 credits, translating to approximately $0.13 per prompt or $0.0325 per generated image variation.

DALL·E 2 doesn’t just rely on simplicity; it packs some robust capabilities. The image editor, albeit in beta, empowers users to include additional generated frames, expanding their images whether they were created with DALL·E 2 or uploaded. This “out-painting” technique allows for the creation of larger AI-generated artistic works.

Furthermore, the tool includes an eraser function, enabling users to eliminate portions of an image and replace them with AI-generated elements—an approach known as “in-painting.”

Moreover, OpenAI offers an API for DALL·E 2, enabling developers to integrate it into their applications. This integration opens up a world of automation possibilities. For instance, you can link DALL·E 2 with Zapier to automate tasks like generating images from Google Forms or HubSpot responses, or you can even use it to streamline your creative inspiration process.

In summary, DALL·E 2’s combination of simplicity, affordability, and useful features makes it a compelling choice for those interested in exploring AI image generation & automation.

DALL·E 2 pros: Incredibly easy to use; relatively affordable—it costs just $0.13 per prompt
DALL·E 2 cons: Pretty basic; generally produces the least photorealistic results; no longer has a free trial.
DALL·E 2 pricing: 115 image credits can be purchased for $15. API pricing is more complex, but starts from $0.016/image.

The AI image generator with the best results

Midjourney

Midjourney consistently stands out as the top choice for generating AI images, producing results that are often the most captivating and visually appealing among all the image generators on this list. Its creations exhibit greater coherence, richer textures, and more vibrant colors. Overall, the output is not only more interesting but also visually appealing.

Notably, Midjourney excels in rendering people and real-world objects, offering a lifelike and natural appearance that requires less extensive prompting. Impressively, it was the first AI image generator to win an art competition, underscoring its artistic capabilities.

However, Midjourney has a significant quirk—it’s only accessible through Discord. Users must join Midjourney’s Discord server or invite the Midjourney bot to their own Discord server. To generate an image, you simply type “/imagine [your desired image]” as a prompt, and the bot provides four variations that can be downloaded, upscaled, and edited.

It’s worth noting that by default, all images generated using Midjourney are publicly posted in the Midjourney Discord community. While this communal aspect is cool, it means that anyone can view your creations. While this might not be an issue for artists, it could be a drawback for business use.

Despite the initial learning curve, Midjourney offers excellent help documentation that guides users through getting started and mastering advanced features. These include using different model versions, upscaling images, blending multiple images, and leveraging various parameters for control. Once you’re familiar with these options, the results you can achieve are truly remarkable.

As for pricing, Midjourney temporarily suspended its free trials due to high demand, although they may occasionally be reinstated for a short period. The Basic Plan starts at $10 per month and includes 3.3 hours of GPU time per month, equivalent to roughly 200 images. Users also have the option to purchase additional GPU time and can use the generated images for commercial purposes.

Midjourney pros: Consistently produces the best looking AI generated images; the community is a great way to get inspiration
Midjourney cons: Can only be used through Discord—which is just strange; images you generate are public by default; free trials are currently suspended
Midjourney pricing: From $10/month for the Basic Plan that allows you to generate ~200 images/month and provides commercial usage rights.

Best AI image generator for customization and control

DreamStudio (Stable Diffusion)

Stable Diffusion is a notable player in the world of AI image generation, and what sets it apart is its open-source nature. This means that individuals with the required technical skills can download and run it on their own computers. It also provides the flexibility to train and fine-tune the model for specific purposes.

Most AI services that generate artistic portraits, historical portraits, architectural renders, and more utilize Stable Diffusion in this manner. If you have the technical expertise and want to create something remarkable with AI, Stable Diffusion is currently the best tool for the job. Additionally, it offers an API and a service called ClipDrop, which breaks it down into a series of tools.

However, Stable Diffusion is also accessible through a powerful public application called DreamStudio, developed by Stability AI (still in beta).

DreamStudio grants users extensive control over various aspects of AI image generation. When you enter your prompt, you can use sliders to determine the final image’s size, how closely it matches your prompt, the number of steps the diffusion model takes, and how many images are generated.

You can also select the algorithm version it uses (the latest being SDXL 0.9) and even input a specific seed for reproducible results (otherwise, they are randomly generated). DreamStudio offers in-painting and out-painting features as well, although you’ll need to use Chrome to access them, and more editing features are on the horizon.

DreamStudio operates on a credit system. Upon sign-up, you receive 25 free credits, which are sufficient for approximately 30 prompts or 120 images with default settings. Using a more potent model, generating larger or more images, or applying more steps will deplete your credits more quickly. When your credits are exhausted, you can purchase more, starting at $10 for 1,000 credits.

In summary, DreamStudio and Stable Diffusion provide unparalleled customization and control over the entire AI image generation process. They empower users to delve as deep into AI as they desire and even build their own AI services. If you happen to exhaust your free trial too quickly, you can explore the same Stable Diffusion models for free through ClipDrop, albeit with watermarked results and less control.

DreamStudio pros: Only major AI picture generator that still offers free credits; incredibly affordable and customizable; super powerful with generally great results
DreamStudio cons: More of a learning curve than some other AI art generators; editing tools aren’t very intuitive.
DreamStudio pricing: Free for 25 credits. From $10 for 1,000 credits.

Best AI image generator for integrating AI-generated images into photos

Adobe Photoshop (Firefly)

Adobe has a rich history of integrating AI tools into its applications for over 15 years, but it’s only recently that the company introduced a text-to-image generator, albeit in beta. You can try it out for free on the web or via Adobe Express, but it truly shines in the latest Photoshop beta, which requires a Creative Cloud subscription to access.

Adobe’s AI model, named Firefly, boasts a range of capabilities. Besides generating new images based on detailed text descriptions, it can create text effects from written prompts (picture the word “TOAST” with letters resembling toast), recolor vector artwork, or incorporate AI-generated elements into your images. While you can experiment with all these features through the web app, it’s in this last aspect that Firefly truly excels.

When evaluated solely as a text-to-image generator, Firefly’s results can be a bit hit or miss. While it can outperform or match DALL·E 2 or Stable Diffusion for some prompts, there may be occasions where its output leaves you questioning its aim. However, its integration with Photoshop, the industry-standard image editor, is where it truly shines.

One noteworthy feature is called Generative Fill. Here’s how it works: You use Photoshop’s regular tools to select an area of your image, and with a simple click and a text prompt, you can replace that selected area with something else. What sets Generative Fill apart is its understanding of your image’s context.

For instance, in the screenshot provided, you can see that Photoshop has maintained the depth-of-field blur and colors to make the castle added via Generative Fill blend seamlessly into the scene.

While DALL·E 2 and Stable Diffusion have initiated discussions about AI image generation, Adobe’s Firefly represents the first implementation of an AI photo generator that offers a glimpse of what the future holds. It’s not just a novelty; it’s a tool that will soon be available to the millions of professionals who rely on Adobe apps daily.

However, it’s important to note that while Firefly is in beta, the images it generates are not intended for commercial use.

Firefly pricing: Free beta online; Photoshop is available from $19.99/month as part of the Creative Cloud Photography Plan.

What about all the other AI image generators?

You’ve made a deliberate choice to focus on the AI image models themselves rather than the applications built on top of them, which is a valid approach. The four AI image generators you’ve selected offer a comprehensive look at the capabilities and potential of these models.

For those seeking a more extensive list of AI image generators, including those integrated into various tools and services, your comprehensive list of two dozen options provides a broader perspective. Each of these tools may cater to different needs and preferences, offering unique features and approaches to AI image generation. This variety allows users to explore and select the tool that best aligns with their requirements and creative objectives.

By offering both a concise selection of top AI image models and a more extensive list of options, you’ve provided readers with a well-rounded resource to navigate the world of AI image generation and find the right tool for their specific use case.

The legal & ethical implications of AI-generated images

The prevalence of AI-generated images raises important questions about their use and the legal and ethical considerations surrounding them.

Currently, there is a lack of clear legal regulations regarding AI-generated content. This ambiguity extends to both sides—AI-generated content is not clearly protected by copyright, and there are no established rules to safeguard artists whose work was used for AI training. To address this, some AI models, like Firefly, rely on licensed images and public domain content exclusively to avoid copyright issues.

The absence of clear regulations has already resulted in lawsuits, with companies like Stability AI facing legal action from entities like Getty Images and individual artists for the unauthorized use of their images. Additionally, class action lawsuits have been initiated against several AI picture generators.

While using AI-generated images for occasional social media posts or blog hero images may not typically lead to legal issues, developing an entire strategy around AI-generated art can be risky due to the current legal uncertainty.

Another crucial concern is bias in AI-generated content. AI often inherits biases from the data it’s trained on, which can manifest in the form of stereotypes or harmful content. Many people have experienced this firsthand, including yourself during your testing of these apps.

Addressing bias in AI-generated content is a responsibility that falls on humans. It involves reviewing AI-generated content for biases and refining prompts to minimize or eliminate these biases as much as possible.

As AI continues to play a growing role in content creation, it’s essential to have ongoing discussions about legal and ethical guidelines, as well as taking proactive steps to mitigate biases and ensure that AI-generated content respects established ethical standards.

What's next for AI image generators?

AI image generation is indeed a dynamic and rapidly evolving field, and the progress made in just a year is remarkable. The growing user base provides developers with substantial data to train and enhance their models continuously, which means we can anticipate ongoing improvements in AI-generated images.

The future of AI image generation holds exciting prospects. As you mentioned, there are potential releases on the horizon, such as Google’s Imagen and Meta’s Make-A-Scene algorithms. These developments could introduce new and innovative approaches to creating AI-generated images, adding further diversity and capabilities to the field.

With technology advancing at such a rapid pace, it’s safe to say that we have much to look forward to in the world of AI image generation. The creative possibilities are expanding, and it will be fascinating to see how these tools continue to evolve and shape various industries and creative endeavors.