OpenAI and MidJourney promotional artworks

For today’s story, I will be comparing OpenAI’s Dall-E2 and MidJourney AI art generator tool with identical prompts.

I have tried several other art generators out there but these two are the most powerful ones so far. Both are still in closed beta, so if you want to try them out, better get on the waitlist now if you haven’t already.

The comparison that I will be doing here is mainly on the visual quality of the output. It means I won’t take into account the differences in terms of other capabilities. For example, MidJourney is able to generate an image with 1792×1024 while Dall-E2 is stuck at 1024×1024 resolution. Dall-E2 can also take an existing image as an input and generate variations of it, while MidJourney can’t.

Okay, let’s make some art!

Side by Side Comparison

Prompt #1: Animated Pixar movie about two motorcyclists

Animated illustration of two motorcyclists. Generated by Artificial Intelligence.

At first glance, the obvious difference is the style. Dall-E2 went to something a little more cartoony compared to MidJourney’s. The facial feature on the right though looks a little messed up, so I prefer the one’s from Dall-E.

Prompt #2: A head of a robot hooked up with wires to the ceiling, illuminated by colored lights

A head of a robot hooked up with wires to the ceiling, illuminated by colored lights. Generated by Artificial Intelligence.

I was already imagining few possible images while typing the prompt, but got surprised when the results came out. Both were so close to what I imagined. Also, Dall-E2’s face features looks a little more accurate than the competition, so my vote goes to Dall-E2.

Promt #3: Spongebob Squarepants painted by Pablo Picasso, oil on canvas

Spongebob Squarepants painted by Pablo Picasso. Generated by Artificial Intelligence.

Again, Dall-E2’s result looks a lot more like SpongeBob. I don’t know why MidJourney has this habit of making some weird alterations of the face. To make it look abstract, perhaps?

Prompt #4: Intricate sculpture of a skull

Intricate sculpture of a skull. Generated by Artificial Intelligence.

It’s a little difficult to decide who did a better job with this one. Dall-E2’s lights and shadows is more pronounced, but MidJourney’s intricate details and overall design is much cooler to me.

Prompt #5: Bright summer afternoon, flowing river, lush nature, studio Ghibli style

Bright summer afternoon, flowing river, lush nature, studio ghibli style. Generated by Artificial Intelligence.

At this point I noticed that MidJourney has this preference of leaning more towards painted art rather than photorealistic. It’s impressive, but putting the Ghibli style in the prompt made me expect something less realistic. MidJourney did a better job on this one in my opinion.

That’s it. What do you think?

If you have prompt ideas that you want me to feed into the AI tools, please let me know in the comments.

How To Get Access

Both Dall-E2 and MidJourney are still in closed beta. In order to get access, you need to register for invite: Dall-E2 apply here. MidJourney apply here.

It took me 4 weeks to get accepted in MidJourney with a few prompt credits to try the bot in Discord. Dall-E2 took a bit longer, around 2 months before I got the beta invite. But the tool is free with unlimited image credits so I have no qualms or whatever about the delay lol.

Final Thoughts

So which tool is better? It depends on your needs.

If you need a free, but lower resolution result, then Dall-e2 is the way to go. If you need a more detailed, higher resolution result, and willing to shell out a few bucks, then MidJourney is the best choice.

These two are not the only AI art generators out there. I compiled a couple more similar tools in the article below.