Midjourney's first new image model in nearly a year, and the one David Holz calls its most intelligent and coherent. Draft Mode makes iteration roughly ten times faster with conversational and even voice prompting; Omni Reference (--oref) succeeds --cref, carrying a specific person or object into new generations. V7 stills feed the 2025 image-to-video frontier — Kling 2.x, Runway Gen-4, Veo 3 — and land just weeks before Midjourney enters the video race itself.