In a week where headlines were dominated by OpenAI’s release of GPT-4 and Google’s AI ambitions, you’d be forgiven for missing that Midjourney released its newest model, V5.
If you haven’t already tested the new model, then you’re in for a treat. People are shocked by how realistic the Midjourney V5 images look. Here are some side-by-side images of V4 and V5 using the same prompts, so you get an idea of how much it has improved.
V4
V5
V4
V5
What is new with Midjourney V5?
V5 has been in the works for five months and uses significantly different neural architectures and new aesthetic techniques. Here are some of the upgrades:
- Much wider stylistic range
- It is more responsive to your prompts. That means short prompts may not work as well. Instead, use more explicit text about what you want (e.g “cinematic photo with dramatic lighting”)
- Much higher image quality (2x resolution increase)
- More detailed images
- Image prompting works much better (i.e. referencing an image in your prompts)
How to use Midjourney V5
To use Midjourney V5, simply add –v 5 at the end of your prompt OR select “V5” from /settings.
It is worth noting that V5 is currently in alpha testing. That means it is likely to be modified before the full release of V5.
Closing Thoughts
This AI model understands prompts and creates realistic imagery like nothing we have seen before.
The exciting thing is that Midjourney’s progress seems to be exponential. With new models released every few months, prepared to be gobsmacked all over again very soon.
If you want to learn more about how to use Midjourney and make AI art, then make sure you check out this guide on advanced Midjourney prompts.
