Midjourney launches its first video model V1 with 5 to 20-second clips

· 1 min read
Midjourney

Midjourney released its first video model, V1, on 18 June 2025. Any subscriber, which includes about 20 million users, can press “Animate” under a still image—whether generated or uploaded—to receive four 5-second clips. These clips are extendable in 5-second increments up to a total of 20 seconds.

There are two motion modes available: low, which involves subtle movements, and high, which includes larger camera or subject shifts. Motion can either remain automatic or be directed by a text prompt. Creating a clip costs roughly eight image credits, with pricing starting at the regular $10 Starter tier. A slower “video relax” queue is currently being tested for Pro users.

V1 lacks audio and is capped at 1080p/20 seconds, which means it trails behind Runway Gen-4, Luma Dream Machine, or OpenAI’s upcoming Sora in terms of scope. However, it undercuts many rivals on cost. Testers have praised the coherence inherited from Midjourney V6.1, although high-motion scenes may experience flickering.

Early reactions vary, from Phi Hoang’s comment that it is “surpassing expectations” to Reddit users noting gaps when compared to Sora-class realism. Analysts describe the move as a quick entry into a crowded field rather than a finished film tool.

Founded in 2022, the San Francisco lab earned $300 million in 2024 from its Discord-based image service. It now faces a Disney-Universal copyright lawsuit over its training data. Executives describe V1 as a step toward a future “world model” capable of creating explorable 3-D scenes.

Source