Midjourney has launched its first video generation V1 model, a web-based tool that allows users to animate still images into 5-second video clips. This new model marks a significant step toward the company’s broader vision of real-time open-world simulations, which will require the integration of image, video, and 3D models to create dynamic, interactive environments.
V1 works by enabling users to animate images through two options: an automatic animation setting, which generates a motion prompt for basic movement, and a manual animation feature, where users can describe specific actions and camera movements. The system is designed to work with images generated by Midjourney as well as those uploaded from external sources, offering flexibility in video creation.
The model also introduces a unique workflow for animating images. Users can drag images into the prompt bar and mark them as the starting frame, then apply a motion prompt to animate them. V1 includes two settings for motion: low motion, which is suitable for ambient scenes with slow or minimal movement, and high motion, which is better for fast-paced scenes with active camera and subject movement. However, high motion can sometimes result in unintended glitches or errors.
When compared to other AI video generation tools currently on the market, V1 offers a distinct approach. Unlike more established platforms like Runway or DeepBrain, which focus on highly polished, pre-built video assets with complex editing features and audio integration, V1 prioritizes the animation of static images within a specific aesthetic that aligns with Midjourney’s popular image models. While competitors like Veo 3 are known for their real-time video creation with full audio integration and high-quality motion capture, V1 sticks to simpler video outputs with limited motion capabilities, focusing primarily on image-to-video transformations.
Midjourney’s V1 Video Model launch has sparked excitement across creative communities, with users praising its stunning visual consistency and artistic flair, often comparing it favorably to competitors.
AI Artist Koldo Huici commented on X:
Creating animations used to take 3 hours in After Effects. Now with Midjourney, I do it in 3 minutes! I’ll tell you how ridiculously easy it is.
While Gen AI expert Everett World posted:
It’s fantastic to have a new video model, especially since it’s made by Midjourney – it opens up new, unexpected possibilities. Some generations look incredibly natural (anime looks great!). Even though it’s only 480p, I think we’re seeing interesting developments in the AI video space, and I’m so glad we can have fun with this model!
Midjourney plans to continue evolving its video capabilities, with an eye on making real-time, open-world simulations a reality in the near future. For now, the V1 model is available for web use only, and the company is monitoring usage closely to ensure that it can scale its infrastructure to meet demand.
This launch comes in the wake of ongoing legal challenges for the company, including a recent lawsuit from Disney and Universal over alleged copyright infringement. Despite these challenges, Midjourney is focusing on expanding its technology, with V1 seen as a significant step toward achieving the company’s vision for immersive, interactive digital environments.
