Midjourney launches the first new model for treating artificial intelligence for nearly a year

Midjourney, one of the first services generated from the web intelligence, has released the first new model of artificial intelligence in nearly a year.

It is called a name V7The model, which started to start in Alpha around the middle of the East night, comes a week after Openai appeared for the first time in a new photo generator in ChatGPT Viral For its ability to create GHIBLI images. Midjourney’s is not an improved model-at least not official-but however generates aesthetic works, at least the eye of this reporter.

To use it, you will first have to evaluate about 200 photos to create Midjourney’s “customizer” profile if you’re not really not. This profile sings the model to your individual visual preferences; V7 is the first Midjourney to which the customization was operated by default.

Once this task is completed, you are welcomed with a wandering v7 on Midjourney on the web, and if you are a member of the Discord’s Midjournyy server, on Discord Chatbot. In the web application, you can quickly identify the model from the drop -down menu next to the “version” name.

David Hones, CEO of Midjourney, described V7 as “completely different engineering” in A. After x.

“V7 is (…) more intelligent with text claims,” ​​Hols continued to announce the dispute. (1) MAGE claims look great, the image quality is significantly higher with beautiful textures, bodies, hands and things of all species have a remarkable better cohesion on all the details. “

V7 is available in flavors, Turbo and Relax – the first is more expensive for operation – and it runs a new tool called the draft mode that makes pictures at a speed of 10x and a half the cost of standard mode. Low -quality images of standard images, but can be enhanced and re -presented with a click of the mouse.

We now start the alpha test phase of the new V7 image model. It is the smartest, most beautiful and most coherent models so far. Give her a snapshot and expect updates every week or two during the next two months. pic.twitter.com/ogqt0fgiy7

– Midjourney (Midjourney) April 4, 2025

Midjourney’s potential is not yet available for V7, according to HOLZ, including increasing images and restoring. He said that they would arrive in the near future – perhaps within two months.

“This is a completely new model with unique strengths and perhaps some weaknesses,” Holls wrote on Discord. (W) You want to learn from you what is good and bad in, but certainly keep in mind that it may require different patterns of claim. So play a little. “

In my summary test, V7 is reasonably committed to the claims that I gave. I was granted, I had no time for truly Put the form through Ringer.

Midjourney is an extraordinary process. It started in 2022 by Holz, who co -founded PC PRIPHERAL COMPANY MOTION, and did not take it from external funds.

In late 2023, Midjourney was It is said It is expected that about 200 million dollars of revenue will be brought. Recently, the San Francisco -based company said Establishing the hardware team To work on unannounced projects, and continue to train pre -announced models to generate 3D objects.

numerous Judicial cases Midjourney was accused of violating the rights of millions of artists by training artificial intelligence tools on the photos that were scraped from the web without the consent of the creators on the photos.