Midjourney, one of the pioneering AI image-generating services, has introduced its latest AI image model, V7, for the first time in nearly a year. The rollout began in alpha around midnight ET on Thursday, shortly after OpenAI launched a new image generator in ChatGPT that garnered significant attention for its ability to create images in the style of Studio Ghibli.
While Midjourney V7 is not specifically designed for Ghibli-style images, it still produces visually appealing results. This new model aims to be the most advanced and coherent version yet, with expectations for updates to be released every one to two weeks over the next couple of months.
To access V7, users must first create a personalization profile by rating around 200 images, which tailors the model to individual visual preferences. Notably, V7 is the first model from Midjourney to have this personalization feature enabled by default.
Midjourney’s CEO, David Holz, described V7 as featuring a “totally different architecture” that enhances its responsiveness to text prompts and improves overall image quality. The model produces better coherence for intricate details, such as body parts and objects.
V7 is available in two configurations: Turbo, which is more expensive to operate, and Relax. It also introduces a Draft Mode that generates images significantly faster and at lower costs, although with reduced quality compared to the standard mode.
Despite its advancements, some standard features like image upscaling and retexturing are not yet available on V7 but are anticipated within the next two months. Holz encourages users to explore various prompting styles to fully utilize the model’s capabilities.
Midjourney operates uniquely; established in 2022 by Holz, the company has not sought external funding yet expects to generate approximately $200 million in revenue. As it continues to innovate, Midjourney is also facing legal challenges related to the use of artists’ works in training its AI.