Midjourney, one of many earliest AI image-generating companies on the internet, has launched its first new AI picture mannequin in almost a yr.
Dubbed V7, the mannequin, which started rolling out in alpha round midnight Jap on Thursday, comes per week after OpenAI debuted a brand new picture generator in ChatGPT that rapidly went viral for its potential to create Ghibli-style images. Midjourney’s will not be a Ghibli-optimized mannequin — not less than not formally — however nonetheless generates aesthetically pleasing works, not less than to this reporter’s dilettante eye.
To make use of it, you’ll first must price round 200 pictures to construct a Midjourney “personalization” profile should you haven’t already. This profile tunes the mannequin to your particular person visible preferences; V7 is Midjourney’s first to have personalization switched on by default.
When you’ve accomplished that activity, you’ll be greeted with a togglable V7 on Midjourney’s web site and, should you’re a member of Midjourney’s Discord server, on its Discord chatbot. Within the net app, you possibly can rapidly choose the mannequin from the drop-down menu subsequent to the “version” label.
Midjourney CEO David Holz described V7 as a “totally different architecture” in a put up on X.
“V7 is […] much smarter with text prompts,” Holz continued in an announcement on Discord. “[I]mage prompts look fantastic, image quality is noticeably higher with beautiful textures, and bodies, hands, and objects of all kinds have significantly better coherence on all details.”
V7 is obtainable in two flavors, Turbo and Loosen up — the previous of which is extra expensive to run — and powers a brand new instrument referred to as Draft Mode that renders pictures at 10x the pace and half the price of the usual mode. Draft pictures are lower-quality than standard-mode pictures, however may be enhanced and re-rendered with a click on of the mouse.
We’re now starting the alpha-test section of our new V7 picture Mannequin. It’s our smartest, most stunning, most coherent mannequin but. Give it a shot and anticipate updates each week or two for the subsequent two months. pic.twitter.com/Ogqt0fgiY7
— Midjourney (@midjourney) April 4, 2025
Quite a few Midjourney capabilities aren’t accessible but for V7, in response to Holz, together with picture upscaling and retexturing. These will arrive within the close to future, he mentioned — presumably as quickly as inside two months.
“This is an entirely new model with unique strengths and probably a few weaknesses” Holz wrote on Discord. “[W]e want to learn from you what it’s good and bad at, but definitely keep in mind it may require different styles of prompting. So play around a bit.”
In my temporary testing, V7 adhered fairly nicely to the prompts I gave it. Granted, I didn’t have time to actually put the mannequin via the ringer.
Midjourney is an uncommon operation. Began in 2022 by Holz, who co-founded PC peripheral firm Leap Movement, it hasn’t taken a dime of outdoor cash.
In late 2023, Midjourney was reportedly anticipating to usher in round $200 million in income. Just lately, the San Francisco-based agency mentioned that it was establishing a {hardware} staff to work on undisclosed tasks, and it continues to coach beforehand introduced fashions for video and 3D object era.
A number of lawsuits accuse Midjourney of infringing on the rights of tens of millions of artists by coaching AI instruments on pictures scraped from the net with out the consent of the photographs’ creators.
