Midjourney V6 Is Finally Here With Exciting New Features

My fellow AI enthusiasts, the moment we’ve been waiting for is finally here — Midjourney V6 has arrived! As a long-time fan of this insanely powerful AI art generator, I was ecstatic when David Holz announced the alpha release just before Christmas.

You can consider this as their holiday present to all Midjourney users.

What’s new in Midjourney V6?

• V6 can understand prompts more accurately
• V6 can follow longer prompts
• Improved coherence and model knowledge
• Improved image prompting and remix
• Minor text drawing ability (you must write your text in “quotations”)
• Improved upscalers, with both ”subtle” and “creative” modes (increases resolution by 2x)

How to activate Midjourney V6

Head over to the Midjourney server on Discord.

Note: If the link above is expired, please let me know in the comments section so I can replace it with a new one.

Open the Midjourney chat instance and type /settings to bring up the settings menu. Under the default model dropdown, select "Midjourney Model V6 [ALPHA]”. You can play around with the other settings, but I will leave them as defaults for now.

Next, type theimagine keyword to bring up the prompt box where you can describe the image you want to generate.

Here’s the result of the following prompt:

Prompt: A shot of a 32-year-old female, up and coming conservationist in a jungle; athletic with short, curly hair and a warm smile — s 250 — v 6.0

The AI will generate four variations of the result. Pick the one you like to upscale by clicking the U1 to U4 buttons.

In this example, I wanted to upscale the second image. Here is the final result:

Here are more sample photos

Prompt: photo of an adult woman. Your beauty is beyond compare

I was stunned by the realistic facial features and depth in the eyes generated.

A brand new “text” feature

I want to give a highlight to this brand new feature that fans, including me, have been waiting for—it's the text feature. Although the Midjourney team claims that the feature is still "minor," my initial tests proved that it’s already impressive.

Here’s an example:

Prompt: A cup of strawberry yogurt with the word “Delicous” written on the side, sitting on a wooden tabletop. Next to the cup of yogurt is a plate with toast and a glass of orange juice

This is beyond my expectations. I can see how useful this will be for brands and illustrators who want to incorporate text into their designs.

I haven’t compared the results yet to other close competitors like OpenAI’s DALL-E 3, Amazon Titan G1, or Adobe Firefly 2.0. If you’re interested in seeing how these AI image generators compare, let me know in the comments.

New prompting method

Midjourney’s new V6 release features a completely revamped prompting system. Founder David Holz emphasized that users will need to re-learn prompt creation compared to previous versions.

The AI is now much more sensitive to specific words and phrases used in prompts. Holz advised avoiding unhelpful terms like “award winning” or technical specs like “4k” that were commonly used before. Instead, users should be very explicit and literal about what they want the AI to generate.

• For more photographic, realism-focused results, it is suggested using the new --style raw parameter.
• Lower --stylize values may improve prompt understanding, while higher values could produce better aesthetics.
• It may be less vibey but if you are explicit it’s now MUCH better at understanding you.

Pricing

Now let’s talk about pricing.

The cheapest monthly subscription starts at $8 and can reach up to $120 for faster generation speeds.

There are also annual subscriptions that save up to $24.

Midjourney made it clear on Discord that, despite the added features and improved speed, the price remains the same.

We’ve pushed our first update to V6, it should be 2.7x faster now (same price). Enjoy!

Few more things to keep in mind

Right now, the available V6 model on Discord is still an Alpha version. Things will change and improve over the next few weeks, according to Holz.

• V6 is slower / more expensive vs V5, but will get faster as the team tries to optimize.
• V6 is the third model trained from scratch on Midjourney’s AI superclusters.
• To learn how to use V6, join Midjourney’s Discord and chat with other users in the ⁠prompt-chat channel.
• Don’t be a jerk or create images to cause drama.

Final Thoughts

Overall, I am impressed with the upgrades of V6 over the previous V5.2 model. Kudos to the team, the AI image generator I loved since it was released two years ago is still my favorite. The dedication and skill of its creators is evident with each iteration.

Three things I’d love to see implemented in the next version:

• Softer tones for more realistic output images. Something closer to Imagen 2’s samples would be great.
• An improved website for image generation. Currently only paid users with over 10,000 generations can access this.
• API for developers — when’s this going to happen? Hopefully soon.

What do you think about Midjourney V6? Has it met your expectations? I’d love to hear your thoughts.