Midjourney, the San Francisco-based AI research lab, has officially launched V1, its first-ever AI video generation model, enabling users to transform static images into short video animations.

Announced on June 18, the new model allows users to create five-second video clips by either uploading their own images or using AI-generated images from Midjourney. By clicking the “animate” option, users can generate four video versions, each extendable up to 20 seconds.

The V1 model supports two animation styles:

Automatic – where the AI suggests a motion prompt

Manual – where users input specific instructions to control how the image moves and evolves

Users can also pick camera styles:

Low Motion – with slow or stationary camera movements

High Motion – for dynamic, action-packed perspectives

In a statement, Midjourney CEO David Holz hinted at the future potential: “The inevitable destination of this technology is real-time, open-world simulations.”

While sound support hasn’t been confirmed yet, the video generation process uses eight times more GPU time than image generation. Despite that, Midjourney claims it is still 25 times cheaper than existing market options.

The feature is available across all subscription tiers, including free users, under the ‘fast mode’, which allocates limited GPU time each month. Users can also access a ‘relax mode’—currently in beta for Pro users and above—which offers unlimited GPU access, though with longer wait times.

This marks a major step forward in AI-powered creativity, making video animation accessible to a broader range of users, from hobbyists to professionals.