Tried and tested: These AI image generators consistently delivered the best results. Here’s a look at how they work, how much they cost, and how they handled a specific prompt.
The use of AI to generate weird and wonderful imagery is definitely my favorite application of the cutting-edge tech.
I’ve tested out AI writing tools and AI productivity tools, but putting the host of AI image generators through their paces was hands-down the most fun I’ve had in testing tools in ages (and I love testing tools).
As with other AI tools, there’s no shortage of artificial intelligence image generators in the market. I’m sure by the time this article is published, about 35 more will have just launched that I’ll need to test.
By now, we all know that AI image generators can come up with all sorts of fantastical, eclectic imagery. But since you’re reading the Buffer blog, I’m going to be presumptuous and assume you’re either in marketing, a small business owner, or a creator — and you’re probably looking for a tool that can do something you can actually use in your work.
Which is a lot like mine. I write content about creators and influencers for this blog and spend a good chunk of my day hunting through free stock photo sites for photos to use as featured images for these articles. So I compared every tool with the same text prompt for an image I often need to find:
“Generate an image of a young, Gen-Z photo-realistic social media influencer creating content by filming a video with their phone in selfie mode.”
In fairness, this is a tough prompt — AI notoriously struggles with human fingers, teeth, and screens. But there’s a reason the platforms on this list are the most popular and talked-about AI image generators of 2024. Of all the ones I tested, these consistently generated the highest-quality images. (I've included some of the best in this article.)
Here are my favorite AI image generators and how they stack up with one prompt in particular.
Jump to a section:
1. Adobe Firefly
Adobe’s answer to generative AI, Adobe Firefly, is baked into its suite of tools, including Photoshop, but there is a free web version available.
What’s great about Firefly (aside from the quality of the tool) is that the images generated with it are safe for commercial use. “Trained on Adobe Stock images, openly licensed content, and public domain content, Firefly is designed to be safe for commercial use,” Adobe says.
The free web version of Firefly, found at firefly.adobe.com, is remarkably easy to use. When you enter your prompt, you can choose aspect ratio, content type (art or photo), and one of hundreds of trending art style reference images (watercolor, studio photography, you name it!). You can even specify the angle of the image and depth of field (close-ups, wide angles, macro photography, etc.), color and tone (pastel, monochrome, muted, and more), and some really fun special effects.
After that, click ‘Generate,’ and Adobe’s AI technology will deliver four new images for you to choose from. Clicking a drop-down arrow on each image will allow you to generate something similar, use it as a style reference for a new prompt, or edit it in Adobe Express (Adobe’s answer to Canva).
My results with Firefly were consistently excellent, with very few Edward Scissorhands fingers. I found that playing around with the various art styles and effects pre-generated really helped me fine-tune my ask. Another strong point in Firefly’s favor was the realism in the human faces in my AI photo — there was none of that creepy Uncanny Valley effect I got with many of the other tools.
Adobe Firefly pricing: Free for the web version (25 generative credits monthly, images with watermarks). Plans start at $5.74 monthly.
2. Midjourney
Midjourney has long been one of my favorite text-to-image generators, but it took some getting used to. The tool is housed in Discord, which is a strange interface for this kind of work, especially if you haven’t used the voice and text messaging platform before.
To get started with Midjourney, you’ll need to sign up on their website and then log in to their Discord server. I’d suggest poking around all the channels to see how things work before you prompt. From there, go to a “newbie channel,” then type /imagine, followed by your text description prompt.
Your results — along with those of all the other folks using the channel — will appear as a reply from a Midjourney bot within a few seconds. You’ll get four unique images, and be able to resize, zoom in, download, or regenerate by choosing the buttons that appear below.
It can feel a little chaotic with prompts and images coming thick and fast, but it can be quite fun to see what others are asking for and the results they get. It also helps that anything you generate will become available in your gallery on midjourney.com
Still, the learning curve is worth pushing through — my results with Midjourney were consistently high quality, and the AI had no problem generating images that really did look like photos.
In all my various regenerations of my ‘influencer’ prompt, there were no glitched fingers, and the people in the ai-generated artworks didn’t have the creepy Uncanny Valley vibe that is often a result of asking AI to create a photorealistic image.
Midjourney pricing: Plans start at $8 per month (paid annually) for limited generations (200).
3. DALL-E 3
OpenAI’s powerful DALL-E 3 is the new-and-improved version of DALL-E 2, one of the first AI image generators to market (and virality). DALL-E 3 is baked right into ChatGPT 4, which is only available on their Plus plan.
The AI art generator is incredibly simple to use — once you’ve created an account, just head over to chat.openai.com and input your prompt into the message box at the bottom of the page. (Be sure to be explicit about the fact that you need an image generated, as ChatGPT can respond in AI text here, too.)
The AI will whip up a single image based on your prompt. From there, you can ask DALL-E to regenerate, download, or edit specific details of the image by highlighting it with a brush tool and inputting a new prompt (a very handy feature).
I tried several iterations and regenerations of my prompt, and the quality was more highly realistic animation than photograph — all of the influencers looked a little too perfect to be human. Still, the textures and image quality were unparalleled. And DALL-E is best known for its more artistic spin on AI image generation.
DALL-E 3 pricing: Included in ChatGPT’s paid plans, which start at $20 per month.
4. JumpStory
JumpStory has two unique selling points: its Everyday AI model is the “world’s simplest image generator” and is only trained on its own library and data. The company says its images are “100% legal with licenses such as Public Domain and CC0 [Creative Commons].”
I can certainly attest to the first promise. To generate an AI image, you simply head over to jumpstory.com in your browser and sign up — no need to download a separate app or program. Choose an art style (photo, illustration, sketch, icon, and a few more) and aspect ratio (square, portrait, or landscape), then input your prompt.
Within a couple of seconds, JumpStory will return four images based on your input. From there, you can either download the files (in JPG or SVG) or ask the AI to regenerate.
Even though the photo image style is still in beta, I was pleasantly surprised by how good they were. The faces did have the telltale airbrushed quality of AI images, and I spotted one wonky finger, but they were of great quality.
JumpStory’s second promise — the fact that you won’t run into any legal issues when using your images commercially — is tougher to verify. It does help that their pro plan ($33 per month annually) comes with “global insurance.” According to their FAQs: “this insurance keeps you safe when you use images that you found using JumpStory in your next blog article, social media post, etc.” Handy.
JumpStory pricing: Plans start at $23 per month annually.
5. Canva
Many of Canva’s best features are hidden behind its paid plans — but not its free online AI image generator. You can access the tool via canva.com/ai-image-generator (you’ll be required to create or sign into an account before you get access, though).
In a new or existing Canva design, you’ll be asked to prompt the AI and choose from a wide variety of styles, from long-exposure photography to anime. Seconds after you hit ‘enter,’ the AI will generate four images to choose from, which can then be pulled into your design.
I was really impressed with how much the images generated for my influencer prompt looked like photographs — until I looked more closely at the fingers. That said, most of the images I generated were usable.
Having this powerful AI image generator baked into Canva is a huge point in its favor, as many social media creators and marketers will likely already be using the image and video editing tool in some capacity. The power of Canva’s vast library of templates coupled with AI makes this one a force to be reckoned with.
While Canva’s AI image generator tool is available on its free plan, users are limited to 50 AI credits. Each change of style or regeneration costs one credit.
Canva AI Image Generator pricing: 50 credits available on Canva’s free plan.
6. Meta AI
Meta (the company that owns Instagram, Facebook, and WhatsApp) launched its own AI chatbot in April 2024, which is accessible on all its platforms. They rolled out the feature to a select group of countries in Africa, as well as Canada, Australia, Jamaica, Singapore, and Pakistan, with more to come soon.
“Thanks to our latest advances with Meta Llama 3, we believe Meta AI is now the most intelligent AI assistant you can use for free,” the company said. “You can use Meta AI in feed, chats, search, and more across our apps to get things done and access real-time information.”
It’s really the company’s answer to ChatGPT, with the added bonus that you can use it right in your favorite social media app. It can do everything that GPT can do — help you create a social media content calendar or plan dinner with whatever you have in your fridge.
While most of the chatter I’ve seen about the tool concerns queries about how to grow your Instagram following (fair play), it features a completely free AI image generator. The downside: all the images generated will have the Meta AI logo and a watermark that reads “imagined with AI.” Meta doesn’t explicitly forbid the use of these images commercially, but the watermark suggests you might want to tread carefully here.
To generate images with Meta AI, head over to Facebook, Instagram, or WhatsApp (you’ll find Meta AI via the search bar) or navigate to meta.ai, then click Imagine. Type your prompt in, then hit enter. In-app, the AI will return one image, whereas using meta.ai will return four. From there, you’ll have the option to edit (via text prompt), animate, and download your AI-generated art.
While Meta AI didn’t generate the best images for my influencer prompt (it had some trouble with photorealism and a couple of bizarre-looking fingers and phone screens), the animated versions (seconds-long GIFs) were really impressive. They’re particularly fun for social media posts.
Meta AI pricing: Free.
7. DreamStudio
Stable Diffusion’s DreamStudio is an AI image generator that offers the user a lot of control over the output (with a handy 25 free credits, too).
Before you hit generate on your prompt, you can choose your image art style (there are 16 in total, which include interesting options like comic book and ‘neon punk’), how many images will be generated, and even how many generation steps will be used.
There’s also a ‘negative prompt’ input box, where you can include things you want to avoid in your AI art, and an upload function, where the AI will create variations of what you’ve uploaded.
Once you’ve inputted your request, you’ll receive one to ten images, depending on the output you selected at the outset. If you’ve found one you like but want to experiment with it, you can select ‘variations’ to have DreamStudio generate more like it.
My mileage with this tool varied. While most of them featured some flaw (those pesky teeth and fingers again), the images generated were among the most visually beautiful at a glance.
DreamStudio pricing: 25 credits for free. Pay $10 per 1,000 credits thereafter (enough for about 5,000 images).
What are the rules around AI-generated images and copyright?
According to the U.S. Copyright Office, AI-generated content isn't copyright-protected, but this doesn’t mean that copyrighted works haven’t been used to train the various AI models these tools use (and may result in a near-replica of another person’s work).
Some of the tools on this list — like JumpStory and Firefly — have been exclusively trained on copyright-free images, but most of them put the onus on the end-user when it comes to legal issues.
“Users get full usage rights to commercialize the images they create with DALL-E, including the right to reprint, sell, and merchandise,” OpenAI says of their tool, and other platforms have similar guidelines. However, all of them absolve themselves of liability for what users create with their tools — so, should there be a legal issue with any AI artworks you use — any ramifications will likely fall on the you.
For each of the tools in question, I’d recommend taking a close look at all the Ts&Cs and fine print that comes with the commercial use of AI-generated art.
Recommended Story For You :
Organize Keywords and Import CSV Files from the Google Keyword Planner
The Most Affordable And Easiest User Friendly Page Builder You Will Ever Use!
Instant WordPress Theme That Matches Your Website
Wizard to Manage Remote Backend MS Access Database Tables Fields and Indexes
If you had an aisle-by-aisle grocery list wouldn't you spend less money on impulse items?
everything you need to create a professional corporate look mini-site is there.
Unlock Your Networking Potential with GNS3Vault
Viper Cache Was 77% Faster Than The Competetion
Understanding Stock Market Shorting eBook
