New Midjourney’s algorithm V4 is out now for testing!
This is a game-changer!
I have been waiting for this new algorithm for weeks! Finally, the exciting message came in the announcements on the Midjourney Discord server:
“Hi @everyone we’re going to try alpha-testing a version of our V4 algorithms today.
To turn it on type — v 4 after your prompt or select “V4” from /settings
What’s new with the V4 base model?
- Vastly more knowledge (of creatures, places, and more)
- Much better at getting small details right (in all situations)
- Handles more complex prompting (with multiple levels of detail)
- Better with multi-object / multi-character scenes
- Supports advanced functionality like image prompting and multi-prompts
- Supports — chaos arg (set it from 0 to 100) to control the variety of image grids
What’s new with the V4 upsampler?
- New upsampler adds/changes/corrects details in the original image (more like the v3 upscaler)
- New upsampler is an early work in progress (it will change significantly)
What are we working on still?
- We will increase the resolution, quality, and controllability of upscaling
- We will add custom aspect ratios (currently limited to 1:1)
- We will continue to increase image sharpness and reduce instances of tight cropping
- Greatly reducing text artifacts (certain prompts become unreadable language on images)
- We’re going to improve the default styling
Please note:
This is an alpha test. Do not rely on any specific quirks or consistency in V4 yet. We’ll be pushing significant changes as we move towards a full launch.
Community Standards:
— This model can generate much more realistic imagery than anything we’ve released before. — From now on, we will be much more strict with our rules around “don’t be a jerk or create images to cause drama”
More about V4:
V4 is an entirely new codebase and totally new AI architecture. It’s our first model trained on a new Midjourney AI supercluster and has been in the works for over 9 months.
V4 isn’t the final step, but our first step, and we hope you all feel it as the new beginning of something deep and unfathomable. Please use these incredible powers with joy, wonder, responsibility, and respect. We’re genuinely excited to see what everyone creates. Have fun!”
If you like my content and want to learn more about AI art and how to make money with it, consider joining Medium as a member through my link. You will be able to read every story on Medium. Thank you so much!
My first tests with Midjourney V4
Of course, I started immediately to test the new version and will write about my experiments in future articles. But for now, I’ll show you my first results and give you my text prompts.
text prompt: beautiful sleeping woman, dreaming of roses, sunshine and angels
text prompt: two cute german Shephard puppies playing with each other on the beach, happy, sunshine
Midjourney gave me three images as an illustration and the fourth was more photorealistic in one batch.
text prompt: big green dragon flying over an old village, spitting fire, dangerous, dramatic sky, cinematic light
text prompt: follow the white rabbit
text prompt: cool dog with sunglasses and a t-shirt, black background, 8K, intricate details
I create designs on request
The first customer of my 5 Stars Designer Membership asked me to create more patterns for print-on-demand and also transparent PNGs with cool and funny-looking animals. He would love to make T-shirts with them in his print-on-demand shop. I will include them in the bonus section of my image library, which you can check out here:
I hope, you enjoyed this short introduction to Midjourney's new algorithm. Will you try it out?
Much love,
Stella Sky