Midjourney launches its first new AI image model for nearly a year

Midjourney, one of the earliest image generation services on the network, has launched its first new AI Image model for almost a year.

Called V7The model that began to unfold in Alpha around midnight east on Friday comes a week after Openai debuts a new image generator in Chatgpt that Quickly passed viral For your ability to create Ghibli -style photos. Midjourney is not optimized by Ghibli-optimized model-at-no-official ones, nevertheless generates aesthetically pleasing works, at least on Dilettante Eye on this reporter.

To use it, you will first need to evaluate about 200 images to build a Midjourney customization profile if you haven’t already done so. This profile adjusts the model of your individual visual preferences; The V7 is the first in Midjourney to include the default customization.

After completing this task, you will be welcomed with Togglable V7 on the Midjourney website and, if you are a member of the Midjourney Discord server, in its discord. In the web application you can quickly select the model from the drop -down menu to the “version” label.

Midjourney CEO David Holz described V7 as “completely different architecture” in a PostS

“V7 is (…) much more comprehensive with text prompts,” Holt continued in a Discord message. “(I) Wizards prompts look fantastic, the image quality is noticeably higher with beautiful textures, and the bodies, hands and objects of all kinds have a significantly better consistency of all details.”

The V7 comes in two flavors, Turbo and Real – the former of which is more expensive to start – and feeds a new tool called draft mode, which makes images at 10x speed and half the price of standard mode. The draft images are of lower quality than standard mode images, but can be improved and rendering with a single mouse click.

We are now starting the alpha-test phase of our new V7 image model. This is our smartest, most beautiful, most renowned model so far. Give him a shot and expect updates every week or two for the next two months. pic.twitter.com/ogqt0fgiy7

– Midjourney (@Midjourney) April 4, 2025

A number of Midjourney options are not yet available for V7, according to Holz, including an increase in images and retext. They will arrive in the near future, he said – probably immediately in two months.

“This is a whole new model with unique strengths and probably a few weaknesses,” Holz wrote for Discord. “(W) I want you to learn from you what is good and bad, but definitely keep in mind that it may require different prompted styles. So play around.”

In my short testing, the V7 clinging wisely to the prompts I gave him. Of course I didn’t have time to really Put the model through the bell.

Midjourney is an unusual operation. It began in 2022 from Holtz, who co -founded the peripheral company of PC Leap Motion, and he did not take a penny of external money.

At the end of 2023 Midjourney was According to messages Expecting to bring about $ 200 million in revenue. Recently, the company based in San Francisco said it was so Create a hardware team To work on undisclosed projects and he continues to train pre -announced video and 3D generation models.

Several lawsuits Accuse Midjourney of violating the rights of millions of artists by training AI tools for images distorted by the network without the consent of the image creators.