The #1 question people have when they want to make AI images: Should I use Midjourney, DALL-E 3, Adobe Firefly 2 or Stable Diffusion? I’ve spent the past 6 months generating 50,000+ images across all 4 platforms. Here’s my breakdown:
Midjourney Beginners & Professionals
This is arguably the most popular platform. For a basic plan of $10 per month, you can’t go wrong with Midjourney. Nearly all of the images that you generate with MJ will look decent. Even if you don’t really know how to write prompts. There are also lots of features for you to explore. You can zoom, vary, upscale, and edit your images with inpainting. Beginners like Midjourney because it’s very easy to generate beautiful images. Professionals like Midjourney because it makes images with the best aesthetics, mood and atmosphere. If you like art and storytelling, Midjourney is for you.
But it has 3 main weaknesses:
a) The interface is not user-friendly. You have to use it in Discord (for now).
b) You cannot upload and edit your own images.
c) Character consistency is a big issue. It’s hard to generate 2 different pictures of the same person.
Want to learn free courses like Agile, Business Analytics, Big Data, Data Analytics, Data Visualization, Microsoft Excel, Product Management, Scrum, SQL, UI/UX: Click here
Midjourney Vs DALL -E 3
The pics that DALL-E 3 generates usually aren’t as beautiful as Midjourney. But it has 4 big advantages over its competitors.
a) It’s free. You can use it in Bing Image Creator. You get 25 credits to generate images every day. Once you run out, you can still generate them, only slower.
b) It listens to instructions very well. Ask it to generate complicated scenes, specifying multiple characters, their clothes and their environment, and DALL-E will do it for you. Other AI’s won’t listen to you when it gets too complicated.
c) It can generate words. Say you want to make a t-shirt design with your name, written in a cursive font. DALL-E 3 can do it – while the other AIs will just create meaningless text.
d) If you have GPT plus, you can use it directly inside ChatGPT. This is incredibly convenient. You can just tell ChatGPT whatever you want to see, and GPT will write the prompts for you. Overall, DALL-E 3 is very easy to use for the average person. It’s good at making stuff like comics, coloring book images, vectors, logos and t-shirt designs. But it’s bad at making realistic and cinematic images. DALL-E also has stronger censorship than the other options.
Adobe Firefly 2
Firefly beats everyone else at one thing: Making realistic photos. It’s not like Midjourney, which always generates beautiful people and perfect scenes. Instead, Firefly is superb at making everyday photos of everyday objects and people. The images are very detailed. They’re so good that it’s hard to tell them apart from real photos. But Firefly is more expensive than Midjourney. It also struggles at other types of images. If you want to make cinematic, fashion, video game or manga scenes, there are better options.
Stable Diffusion
Stable Diffusion can do almost everything that the others can and more. It can generate 100 pictures at once. It can increase the resolution of a photo by 10x. You can train it to make pictures of your face. There’s an almost unlimited number of styles that you can download from Civitai. SD also has a powerful feature called ControlNet to control the exact pose and style of the images and it has zero censorship. But the thing is, for ordinary people, it’s hard to learn.
If you want to get superior results, you need to spend a lot more time researching models and tweaking parameters. You also need a good GPU with at least 8GB of RAM to run it properly – or rent a cloud GPU. Therefore, if you’re a casual user, there’s no need for Stable Diffusion. But if you’re a pro who wants to train an AI to make pics of your own artwork or a tech geek who loves fine-tuning models or a brand seeking the best shots for your product then Stable Diffusion might be the right choice.
Conclusion
Which AI is best, really depends on what you want to use it for.
If you like cinematic shots and beautiful aesthetics, use Midjourney.
But you want something straightforward with lots of practical use cases, use DALL-E 3.
If you want to make super realistic photos, use Firefly 2.
If you’re not satisfied with the others and wanna go hardcore mode, use Stable Diffusion.

Moringa: Nutrient-Packed superfood, easy to find in Pakistan