Best Free AI Art Generators, Ranked (2023)

The impenetrable hive mind of the internet is a better artist than me. That's not a point of contention, that's a fact. Sure I can write, but you shouldn't witness the slapdash nature of my attempts at oil painting - all those atrocities lingering on a basement shelf. Even the worst AI image generators and theirsstrange obsessionswith certain wardrobes have an odd—almost disturbing—flowing quality that makes each attempted depiction so distinctive.

Aside from the question of whether AI-generated images are legitimate "art," the best system-produced images are more than just means to create weird nightmaresDepictions of celebrities in different art styles. At the same time, it is not enough for the digital artist to vaguely offer an impressionistic, oddly shaped "interpretation" of the original images or user prompts.

In all this jostling between technology and art, there is a middle ground that even the most untrained in the art of painting on canvas can understand. What can inspire us? What can fascinate us? This is what AI image generators are capable of.

So we turn to the free AI art generators, or at least the ones that offer free trial options. To best classify each program, I've given them all the same, rather esoteric, textual prompts based on some books I've read recently. These books include:

Under the pendulum sunby Jeanette Ng

Prompt: "A man and woman stand under a pendulum sun in the heart of Arcadia."

The dispossessedvon Ursula K. LeGuin

Prompt: "A lonely mathematician stands on a dusty planet and owns nothing."

A memory called Empireby Arkadi Martine

Prompt: "A strange woman fights alone against the machinations of a cosmic empire."

There are several image generators that require photos instead of text prompts. As much as I want to keep it consistent, I want to include different systems instead of excluding them. For the image generation platforms that don't allow text prompts, I used the same image for each:

I'm not an art critic, but at least I can tell if AI generated art actually attempted to represent a prompt in a way that wasn't derived or relied uponCopying and Replicating Artfound in the Internet.

To be honest, I'm pleasantly surprised by some of the results from some of the more popular free art generators. let's see, shall we?

Oh boy turn your image into NFT "art" what a great design. Well, apart from trying to monetize the poor greyhound, let's see whatFotorcan do to make Skip a little more picturesque.

With the program's GoArt creator, there's honestly nothing here that I haven't already seen from a Photoshop filter. While the program is relatively easy to use, I don't really see the fun in making an image a little more sketchy. The program does have a nice little photo editor, but if you want an image without a watermark you'll have to pay, but you really shouldn't expect anything less. There are better free programs available online if that's all you're looking for.

stewis a fairly rudimentary program, but it allows users to create relatively simple AI art at a reasonable price. Although in the end it doesn't really understand what to do. It's hard to see the original prompts in anything the AI ​​system can produce. Like many image generators, it struggles with faces.

DeepAIsThe rather basic and rudimentary AI generator doesn't have a lot of bells and whistles and despite that... doesn't have much else to offer anyway. The text-to-image API system just doesn't have the bells and whistles that more modern systems have. The images are more collages of images found on the internet than a genuine attempt to create something "new".

Thatdream studiobeta AI art generator is free and has a lot of bells and whistles that other AI art generators don't have, like the ability to scale the width and height of the image while also telling you how close you want the image to be to your prompt. It was developed by Stability AI, who recently released a much more popular AI art generator (click through to learn more about Stable Diffusion).

The system is also surprisingly fast, just from the examples above you can see that the AI ​​isn't afraid to mix and match art styles at whim.

Unfortunately, using my prompts resulted in some interesting renditions, but also some weird and weirdly derived images. Several times when I used itA memory called EmpireI promptly received images with broken text in no human language, almost like an alien book cover. It's a sign that the system really wants me to give it more information, tell it to rip off an artist or two. It's much better if you give it the name of an artist to copy from, but in that respect it's similar to our next AI art generator.

Much has been said about stability AIFree art generatorwhile it was in closed beta but on22. AugustThe free image generator finally got a full release and has been one of the most talked about image generators ever since. Stable Diffusion is open source, free and unfiltered compared to Dall-E or Google's (still not publicly available) Imagen. The basic side of Hugging Face is pretty rudimentary, but you can make it a lot easier to use with something simpleFreeware-Tools, although you need to install Python for it to work. Stable Diffusion relies on a model that is based on the LAION-5B data set and, according to its own website, filters out images and logos with watermarks. Based on a report byTech-Blogger Andy Baio, many of the images are mainly sourced from Pinterest and other photo and art blogs. My attempts to ask it to create its own art without telling it to copy a specific artist have been largely unsuccessful. It kept offering me black and white images without any kind of style or substance. Without asking it to copy any particular artist, it just offers pretty disappointing works compared to whatother usersmanaged to get.

It seems that the system is much better at developing art if you specifically ask it to do it in a specific artist's style. This, of course, leads to a host of ethical issues, particularly for the living artists, which people may emulate, as demonstrated by fantasy artist Greg Rutkowski, who was interviewed byMIT Technology Reviewand said he was concerned that the number of fake AI artworks bearing his name would eclipse his own visibility.

You must love freeware, and this one is Python basedGoogle CoNotebook is relatively easy to use, so you really have to make big props for a system that's both relatively simple and open and available to all users. All you have to do is go to the link, scroll down and type your command prompt in the text box, then either press Ctrl+F9 or Runtime - Run All.

But of course we are the artistry of the system. The system iterates over the design step by step so you can see where the AI ​​is trying to go with each frame. My prompts offered some really interesting results, but also a few head-scratchers.

The system that was once known as the Dall-E Mini is nowcrayon, creates a variety of different images that it hopes will fit the bill, so from all of them I've picked the one that comes closest to the spirit of the prompt. The images themselves are quite low resolution, which detracts from the fact that you have so many options to choose from based on a single sentence.

The art also left a lot to be desired when it came to choosing between nine different tiles. Compared to other AI art, Craiyon really feels like it's collecting images from all over the internet and turning them into a hodgepodge of an approximation of what it believes to be the prompt.

GooglesDeep Dream Generatorclaims that it can transform images into stylized works of art. The main feed of images on the site makes it look very impressive, but of course it's not exactly easy to make an image of young Skip appear as anything more than just a dog with the Photoshop filter on.

I have young Skip in the style of Pieter Bruegel, the famous Dutch artist behind works likeThe blind lead the blind.It came out... okay. I tried a mandala pattern and it was...interesting I think. Then I added a Salvador Dali painting to the style and – well – it came out kind of boring.

It's definitely a fun tool, but it feels like I'm playing with an elaborate coloring book instead of creating real art.

ThatNight cafeThe system only lets you take a few pictures before asking for payment, but like some other AI picture generators, it gives users a whole range of different styles to choose from. It wants users to buy "credits" to make more art or improve the quality and resolution of each image, and you'll run out pretty quickly.

The art itself is everywhere. I appreciate what it was trying to do with itThe dispossessedbut it wouldn't even spawn a human-looking creature in either of the other two prompts. The image resolution isn't great either, and the art is very strange even on the default oil painting setting. Sure, it's nice to be able to try out the system to see if you like it at first, but even at just $10 a month I can't say you won't find better options out there.

I guessStarryAIsclean UI and simple systems, although of course this is another one that gives users a few free credits to start with before eventually asking you to cough up more. Again, there's nothing wrong with making users pay to create a whole bunch of art, but it has to be worth it. This art generator also allows users to upload an initial image to give the AI ​​a head start, but no hand-holding is allowed here for these leaderboards.

The art itself is a mixed bag. I like what it came up withThe dispossessedbut can't really say what it had in mindA memory called Empire.

DALL-E 2 may have been the most anticipated AI image generator since it first entered its beta version in April. However, as these examples show, there have been several releases to capitalize on the AI ​​art craze.

Finally OpenAI, the makers of DALL-Epublished their systemto the general public on September 28th. Like many other systems, it offers users 50 free prompts, plus a few more every month. Users can also pay for more credits for image generation.

My experience with DALL-E was interesting to say the least. I find that the system doesn't calculate certain artists all that well, and even if you give it fairly detailed instructions, it defaults to an Impressionist art style. Of course, that is not the aim of this ranking. I'm more interested in what each AI system can do on its own without giving it an image that it can specifically replicate. DALL-E's systems obviously draw from a fair amount of real photos and artwork, but it still largely fails to replicate faces. However, the generator seemed to like the prompt forA memory called Empireand it offered several inspirational images that capture the essence of the book.

ThatContinuation of the first stable diffusion, developed by Stability AI, may call itself a full iteration compared to its original incarnation, but the new system, released at the end of November, is certainly an upgrade of its predecessor in many ways.

In hisannouncement blog, Stability AI found that it had introduced a new text encoder developed by LAION, the open-source project that also provided the vast set of images that Stable Diffusion draws from.Stable Diffusion 2also displays images with a much better resolution and, according to Stability AI, can also upscale them better. They also mention that it offers a greater ability to facilitate depth of field, although in my own testing with my command prompts that didn't really show up.

When I used the system to generate my open-ended prompts, I was surprised by the sharpness of some of the images generated. I want the AI ​​to generate images without specifying a specific artist, style, or format. Without causing it to base a face on a person, the system still struggles to produce a realistic head. Still, I was surprised at what it brought out for meUnder the pendulum sunprompt.

Weirdly, I fed Stable Diffusion 2 every time I mineA memory called EmpireA comic-inspired display promptly followed. I picked the best one, but what's of particular concern is that the system still wants to replicate text. It's further evidence of how much the system likely borrowed from scans of actual comics and comic artists.

ThatWombo DreamThe system allows you to create art in several different styles, e.g. B. old retro art, Salvador Dahli or - simply - "Ghibli". I chose a different style for each based on the style of each book. It also allows you to insert a reference image that Dream can use, but I've limited the system to its own imagination. You can also convert the images to NFTs, but thanks, no.

I have to give it to Wombo, some of these artworks are really impressive. I'm particularly drawn to the Throwback filter as it definitely gave it that 1970s art style you might see on an old album cover. However, as you can see, it is very loose with its interpretations. The oddly erotic feeling I get from his interpretationA memory called Empireis strange since that wasn't the prompt. I was very surprised by his interpretation ofunder the pendulum sun,so it gets pretty good grades.

The crown of the AI ​​art generation scene falls on none other thanOccasionally. It's gotten quite a bit of hype since its open beta first burst out of the Discord server weeds, but it's honestly one of the best AI art generators available. Although I've seen some rather strange renderings of ships and other objects, the sharpness of the images produced might convince someone they were man-made.

I'll say that Midjourney has a bit of a sense of repetition, as I've seen similar poses in the stoic, feminine design among other user-driven images. That doesn't take away from how impressive the art is, though, especially based on such weird, convoluted prompts. I mean look at it! Some of the images are based onThe dispossessedPrompt could be the actual cover of the book. Some of these in the last panel are slightly as I would imagine Mahit DzmareA memory called Empirewould look like.

The program remains committed to Discord, and the bot offers the first generated images as a trial, but then asks users to pay $10 for a basic monthly membership with 200 images.