Midjourney has announced the launch of its highly anticipated Video Model V1, marking a significant step toward the future of real-time, open-world AI simulations.

From Still Images to Dynamic Worlds

For years, Midjourney has been at the forefront of AI-generated imagery. Now, the company is setting its sights on an even more ambitious goal: creating models capable of generating real-time, interactive 3D simulations. Imagine commanding an AI to generate visuals that move and respond in real time—allowing users to explore, interact, and even direct the action within fully realized digital environments.

To achieve this vision, Midjourney is methodically building the necessary components:

• Image Models: The foundation, enabling high-quality visuals.
• Video Models: Bringing those images to life with movement.
• 3D Models: Allowing navigation through space.
• Real-Time Performance: Ensuring everything happens instantly and seamlessly.

The company plans to develop and release each building block over the next year, gradually integrating them into a unified system. While early adoption may come with higher costs, Midjourney anticipates rapid improvements in accessibility and affordability.

Introducing the Image-to-Video Workflow

Today’s major announcement is the rollout of the new “Image-to-Video” workflow. Users can now transform their static images into animated sequences with just a click. After creating an image in Midjourney as usual, simply press “Animate” to bring it to life.

Key features include

• Automatic Animation: Instantly animates images using AI-generated motion prompts for effortless creativity.
• Manual Animation: Gives users precise control, allowing them to describe exactly how they want scenes and objects to move.

Motion Settings

Low Motion is ideal for subtle, ambient scenes with minimal camera movement.

High Motion is suited for dynamic sequences where both the subject and camera are in motion, though it may occasionally produce unexpected results.

Videos can be extended in four-second increments, up to a total of four extensions per animation. Users can also animate images uploaded from outside Midjourney by dragging them into the prompt bar, designating a start frame, and specifying the desired motion.

Accessible Pricing and Responsible Use

At launch, the Video Model is available exclusively via the web. Each video job costs roughly eight times more than an image job, but produces four five-second videos—making the price per second comparable to a single image upscale. This pricing model is over 25 times cheaper than previous market offerings and is expected to become even more affordable as the technology matures.

A “video relax mode” is also being piloted for Pro subscribers and above, offering additional flexibility for power users.

Midjourney emphasizes the importance of responsible use, encouraging the community to explore the technology’s creative and practical potential while remaining mindful of its impact.

Looking Ahead: The Road to Real-Time AI Worlds

This release is just the beginning. Insights gained from building the Video Model will soon inform improvements to Midjourney’s image models and pave the way for future advancements in 3D and real-time simulation. The company is committed to refining its offerings, monitoring usage, and ensuring a sustainable business model as demand grows.

Thank you for being part of this journey with us—and have fun!

David, Midjourney Team

Midjourney’s latest innovation signals a new chapter in AI-driven creativity, bringing us closer to a future where anyone can generate and explore living, breathing virtual worlds in real time