Who Wins in Image Generation? Midjourney vs ChatGPT: 5 Examples

I wanted to see firsthand how Midjourney and ChatGPT stack up against each other in image creation, so I put them to the test with five unique examples, giving us a clear picture of which one shines brighter for our creative needs.

Image Prompt Example 1: Snapchat Train

💡 Midjourney version 6:
/imagine snapchat image of a person sitting in train

Here’s an image created by Midjourney:

💡 GPT-4 (ChatGPT)
create image: snapchat image of a person sitting in train

Here’s an image created by ChatGPT:

⚽ Current Score: Midjourney 1: ChatGPT 0

Image Prompt Example 2: Soccer Lego Lady

💡 Midjourney version 6:
/imagine female soccer player (lego)

Here’s an image created by Midjourney:

💡 GPT-4 (ChatGPT)
create image: female soccer player (lego)

Here’s an image created by ChatGPT:

Both are great images, so I’ll give a +1 for both models: 👇

⚽ Current Score: Midjourney 2: ChatGPT 1

Image Prompt Example 3: Klimt Hacker

💡 Midjourney version 6:
/imagine painting of a programmer with hoodie in the style of Klimt

Here’s an image created by Midjourney:

💡 GPT-4 (ChatGPT)
create image: painting of a programmer with hoodie in the style of Klimt

Here’s an image created by ChatGPT:

Although the ChatGPT image has elements of Klimt, the programmer himself doesn’t use the “Klimt style” at all, so the overall image is flawed. Midjourney respected the request much better:

⚽ Current Score: Midjourney 3: ChatGPT 1

Image Prompt Example 4: T-Bone Steak Shuttle

💡 Midjourney version 6:
/imagine a tasty T-Bone steak flying through space with microbacteria living on the T-Bone space shuttle --v 6.0

Here’s an image created by Midjourney:

💡 GPT-4 (ChatGPT)
create image: a tasty T-Bone steak flying through space with microbacteria living on the T-Bone space shuttle

Here’s an image created by ChatGPT:

Although the Midjourney version looks slightly more spacy, ChatGPT respects my prompt better:

⚽ Current Score: Midjourney 3: ChatGPT 2

Image Prompt Example 5: Broccoli Train

💡 Midjourney version 6:
/imagine feature rich child image of a broccoli train driving through milk galaxy

Here’s an image created by Midjourney:

💡 GPT-4 (ChatGPT)
create image: feature rich child image of a broccoli train driving through milk galaxy

Here’s an image created by ChatGPT:

ChatGPT wins this one — getting even the green/white milk/broccoli mix right. Also the train itself partially consists of broccoli.

⚽ Current Score: Midjourney 3: ChatGPT 3

Surprisingly, there was no clear winner. I expected Midjourney to win this battle but subjectively, I cannot really see this confirmed with my little experiment.

Also check out this article:

💡 Midjourney V5 vs V6: 10 Portrait Examples That Show the Differences (Before/After)