Are you ready to take your image output to the next level? Midjourney V5 has arrived, and it's packed with new features and improvements. In this article, we'll explore what's new in V5, compare it to previous versions, and showcase some sample prompts and outputs to demonstrate the evolution of image output quality.

What's New in Midjourney V5?

Let's start with the official version of what's new in Midjourney V5, and later on, we'll mention our unofficial findings on what's cool.

--tile Argument

The --tile argument is back in V5, which is used when creating a pattern. This argument was temporarily taken out in V4, but it's now reintroduced in V5. Here's an example:

Midjourney Tile example

/imagine MYPROMPT, repeat, white and blue highlights --s 1000 --v 5 --tile

--ar Argument

In V4, custom aspect ratios were limited to those that were either not taller than 1:2 or not wider than 2:1. This limitation meant that you couldn't have a ratio like 4:1 or 1:3, making cinematic widescreen images impossible to generate in V4.

In V5, you can create your own custom ratios with any width:length ratio you want! That means you can now generate cinematic renditions of images, like anamorphic frame formats with ratios of 2.39:1, which are used in artistic/cinematic creations. Here are some examples of custom aspect ratios:

Different aspect ratios supported by Midjourney

--iw Argument

In V5, you can now customize the weights from image prompts. This feature allows you to decide how much influence the input image has on the final generated image. Here's an example:

Passing an image as input along with the text prompt in Midjourney

Comparison of difference in image weights (Image generated in Midjourney)

/imagine a roman soldier with daisies in the background --seed 123 --iw 2 --v 5

V5 is "Less Opinionated"

Unlike V4, which generated more creative output images from smaller text prompts, V5 requires more adjectives and details on the lighting, style, or mood of the output image. If you don't specify these details, the default style assumed is 'Photographic.'

V5 offers more control over output generation through text prompts, resulting in greater cohesion between the input text and the output image.

Unofficial Findings

We've also made some unofficial findings on what's cool about Midjourney V5. Here are a couple of improvements:

  • Fingers and toes are rendered almost to near perfection.
  • Window reflections and photoshoots through glass/windows look more realistic.

Realistic Environment

V5 moves from a more studio-like environment in V4 to a more realistic environment, as long as no text prompts are provided about the environment.

The output images in V5 are more vibrant and offer a wider stylized option than V4.