Midjourney, one of the earlyest AI image creation services on the web, has launched its first new AI image model in almost a year.
The model called Dub, called the V7, began to be released in Alpha at midnight on Friday, and it quickly reached the virus with the ability to create a Ghibli -style photo after a week after the Openai debuted a new image generator in CHATGPT. Midjourney’s S is not at least officially not officially a Ghibli-optimized model, but nevertheless creates a aesthetically aesthetic work for the reporter’s dillettante.
To use this, you should first evaluate about 200 images and do not yet see the mid -jay “Personalization” profile yet. This profile adjusts the model to individual visual preference. The V7 is the first Midjourney that turns on personalization by default.
If you complete the task, you will find a Discord Chatbot if you are a Discord Server member of the Midjourney website. In the web app, you can quickly select the model from the drop -down menu next to the “version” label.
Midjourney CEO David Holz described the V7 as “completely different architecture.”
Holz said in a presentation on Discord, “V7 is much smarter to the text prompt.” (I) The wizard prompt looks fantastic, and the beautiful texture makes the image quality noticeably higher, and all kinds of body, hands and objects have a much better consistency in all details. “
The V7 is provided with two flavors of turbo and rest, of which the former costs more. It also delivers a new tool called draft mode that renders 10 times the speed and half of the standard mode. The draft image is lower than the standard mode image, but it can be enhanced and rendered with a mouse click.
We have now started the alpha test stage of the new V7 image model. It is still the smartest, most beautiful and most consistent model. Expect update every week or two for the next two months. pic.twitter.com/ogqt0fgiy7
-Midjourney (@midjourney) April 4, 2025
According to HOLZ, V7, including image up scaling and re -text, is not yet available. He said he would arrive in the near future.
Holz about Discord, “This is a completely new model with unique strengths and perhaps a few weaknesses.” (W) E wants to learn from you, but it can definitely need a different style of prompt. So please play a little. “
In my simple test, the V7 was reasonably good at the prompt I gave. Of course, I had no time really Put the model through the ringer.
Midjourney is an unusual operation. HOLZ, which co -founded Leap Motion around the PC, started in 2022 and never paid outside money.
In the late 2023, Mid Gerry expects to generate $ 200 million in sales. Recently, the company, headquartered in San Francisco, is establishing a hardware team to carry out unveiled projects and continues to educate previously released video and 3D object creation models.
Some lawsuits accuse Midjourney for infringing the rights of millions of artists by training AI tools for scratched images on the web without the consent of the image producer.
