Midjourney 5.1 vs. Midjourney 5

Enhanced image system with improved coherence and sharpness

By Matic Broz, the editor-in-chief with 10+ years of experience with design, stock media and licensing, and photography.

Updated on | 0 Comments

If you buy something from a Photutorial link, we may earn a commission. See our ethics statement.
Adorable teddybear Midjourney 5.1

Midjourney has just announced the release of their updated image system, Version 5.1. With this new version, users can expect a more opinionated approach similar to Version 4, offering a more user-friendly experience with short prompts. Additionally, an ‘unopinionated’ mode called “RAW Mode” has been introduced for those who prefer the default settings of Version 5.0. The company reassures that Version 5 will still be available for use.

Major improvements in Version 5.1 include:

  • higher coherence,
  • increased accuracy to text prompts,
  • fewer unwanted borders or text artifacts, and
  • enhanced sharpness

To enable Version 5.1, users can simply navigate to /settings and click “V5.1” or type–v 5.1 after their prompt. For RAW mode, click the RAW Mode button under /settings or type –style raw.

As part of the update, Midjourney is also re-enabling the AI moderation for Version 5 and newer models, with improvements based on user feedback from the previous test run. The company will continue to collect data and refine the system over the next few days.

Testing Midjourney 5 vs. 5.1

I tested the new version of Midjourney by using exactly the same prompts with version 5 and version 5.1 and comparing the results. All other settings were left the same, set to base quality (–q 1), medium style, and I didn’t cherry-pick the results—I used the first grid generated by each model.

In all the images below, the four images on the left are version 5, and the four images on the right are Midjourney 5.1.

First, I tested if Midjourney was able to mend the weird-looking flares. Here, the prompt was “A panoramic view of a mountain range at sunrise, captured in the style of Ansel Adams’ black and white landscape photography, featuring dramatic contrast and sharp details, with a focus on the interplay of light and shadows, using a large-format camera“.

Panoramic view of a mountain range at sunrise
Midjourney 5 (left) vs. 5.1 (right) (by Matic Broz)

As you can see, the shadows and sunrays in the version 5.1 images look more natural, with also details in the foreground. The sun flares aren’t as aggressive in version 5.1 compared to version 5.

The second prompt I used was “Graceful ballerina dancing on stage, captured in the style of impressionism, emphasizing movement and emotion, inspired by the works of Edgar Degas, featuring pastel colors and soft, blurred lines, with a focus on portraying the delicate balance between strength and beauty, using a telephoto lens.“:

Graceful ballerina dancing on stage Midjourney 5 vs 5 1
Midjourney 5 (left) vs. 5.1 (right) (Matic Broz)

Once again, the images in Midjourney 5.1 are much more cohesive than the version 5 equivalents. The background colors in version 5 images are all over the place and don’t follow the prompt “pastel colors” as well as the version 5.1 images.

After seeing this, I was interested in how Midjourney 5.1 would handle expressive portraits and their colors, so there’s the prompt I used, “A candid portrait of a musician performing on stage, captured in a dynamic and energetic way, inspired by the works of Mick Rock and Annie Leibovitz, featuring a mix of wide shots and close-ups, with a focus on the personality and expression, using a fast zoom lens“:

Candid portrait of a musician performing on stage Midjourney 5 vs 5-1
Midjourney 5 (left) vs. 5.1 (right) (by Matic Broz)

Immediately, you can notice that the background colors in version 5.1 images express the prompt “dynamic and energetic” perfectly, while the Midjourney 5 images are a tad boring. Yet, the faces in both images seem equally good.

In my review of the best image generators, I observed that all AI image generators struggle with faces. So, what better way to test it than a group of people? Here’s the prompt, “A group of diverse people protesting for social justice in a city street, captured in a documentary-style editorial image, emphasizing the energy and emotion of the moment, inspired by the works of Gordon Parks, featuring a mix of candid and posed shots, using a fast prime lens“.

Group of diverse people protesting for social justice in a city street Midjourney 5 vs 5-1
Midjourney 5 (left) vs. 5.1 (right) by (Matic Broz)

Midjourney 5 already draws realistic faces, much better so than DALL·E 2. The key difference between Midjourney 5 and 5.1 is that the version 5.1 faces are very expressive and match the attitude of the event, while the version 5 faces are all emotionless.

Otherworldly landscape of a distant planet Midjourney 5 vs 5-1

Midjourney 5 (left) vs. 5.1 (right) (Matic Broz)
Prompt: An otherworldly landscape of a distant planet, depicted in a science – fiction inspired way, featuring surreal colors and textures, with a focus on the sense of awe and wonder, inspired by the works of Chris Foss and Moebius, using a fisheye lens

How to get Midjourney 5.1?

Midjourney is available with Midjourney subscriptions, starting at $8 per month.


Leave a Comment