Midjourney Debuts First Video AI Model to Pave Way for Real-Time Imagery

Midjourney has released its first-ever video generation model, marking a significant shift from still imagery to animated visuals. The V1 Video Model, launched on Thursday, introduces an image-to-video feature that animates user-generated images using automatic or manual prompts.

The feature allows users to click the ‘Animate’ option on any image and choose between low and high motion settings. While low motion suits ambient scenes with subtle movement, high motion adds dynamic camera and subject motion, though with a greater chance of visual glitches. Each animation job generates four five-second clips, which can be extended.

“Once you have a video you like, you can ‘extend’ them—roughly four seconds at a time—four times total,” the blog post read.

The company calls this launch a stepping stone toward its long-term goal of enabling “real-time open-world simulations”, where AI-generated characters and environments are interactive and move naturally in 3D space.

“Our goal is to give you something fun, easy, beautiful, and affordable so that everyone can explore. We think we’ve struck a solid balance. Though many of you will feel a need to upgrade at least one tier for more fast-minutes, the company’s blog post stated.

Users can also animate external images by uploading them and entering a custom motion prompt. Midjourney aims to keep the process simple, creative, and affordable, though it notes the computational cost is about eight times that of a typical image job. That said, each second of video is priced comparably to one image upscale—over 25 times cheaper than similar tools currently available, according to the company.

The new tool is initially available on the web, with pricing and server capacity expected to evolve in the coming weeks. A “relax mode” for video generation will also roll out for Pro subscribers and above.

With this release, Midjourney joins the growing list of generative AI platforms exploring video, alongside Google Veo, OpenAI’s Sora, and Runway. This indicates an accelerating race to bridge static imagery and dynamic storytelling.