Midjourney v6 alpha is here!

It was a long wait, but finally, Midjourney v6 is here.

The responses have been somewhat mixed, from excitement to lukewarm.

Just before the release, there seemed to be a bit of a vibe shift toward Midjourney on X. Did you notice it?

But, we did get v6, so let’s take a look at what it can do.

Midjourney v6

Better prompt adherence.

v6 is able to take more from the prompt that you give it, resulting in more precise output.

It’s still not perfect, especially with a long prompt, but it can handle them pretty well too.

In the above demo, IMO, Midjourney generated all of the elements from the prompt correctly.

Can render text.

Text generation works with v6. However, it’s not 100%. You may need to do a few reruns.

A word and a few words work better than long text blocks.

You need to add any text you want Midjourney to generate in “quotation marks”

What is fantastic is that Midjourney matches the font and the style of the text to your image. This is something Dalle-3 is not able to do.

Better language understading.

Try scorpion, floating, or elf ears - Midjourney can now do those, among other things.

It’s still not perfect, but it’s much improved.

Prompting - is different.

There is going to be a bit of a learning curve to get used to this new style of prompting.

Use more natural language, and be specific to what you want. For example, you many need to add more details, or change a few things.

I did a few test and at least for those also comma separated prompts worked fine.

Some of the v5.2 prompts generate fairly similar results with v6, but other do not - at all.

David also mentioned that lenses should work on v6 prompts too. I have not tested this yet.

You may have noticed that zoom, panning and inpainting are missing. These were not included in the launch and they will be added on later.

I am feverishly prompting old prompts and enjoying exploration with v6.

I’m wishing you the warmest holiday greetings with hugs.

Alie