Midjourney V7 is a completely rebuilt model featuring a redesigned architecture, delivering exceptional prompt comprehension and richer image quality across the board. Powered by a 12B parameter Diffusion Transformer, it brings significant improvements in rendering human anatomy, hands, and object details. Draft Mode enables rapid low-resolution previews for faster iteration, while Omni Reference ensures consistent character and object representation across generations. With personalization built in by default, V7 learns your aesthetic preferences and applies them to every image it creates.
Vision|Proprietary Model