MIDJOURNEY
Midjourney just made your images move, literally
After years of focusing on image generation, the Midjourney team has launched Version 1 of their Video Model, taking the first real step toward real-time, open-world AI simulations.
Think: 3D environments, moving characters, and interactive scenes , all generated by AI, in real time.
But for now, it starts with animating images.
The new tool, Image-to-Video, lets users turn still images into short animations.
You can hit "Animate" and let the system decide the motion (automatic), or guide it yourself with a custom prompt (manual).
There’s a choice between “low motion” for ambient scenes and “high motion” for more dynamic ones, both with their own quirks.
Users can extend their videos in 4-second increments (up to 20 seconds total) and even animate images uploaded from outside Midjourney.
What to know:
You can now animate images in Midjourney with auto or manual prompts
Pricing starts at one image per second of video, much cheaper than the competition
Real-time, open-world AI simulations are the long-term goal
We’re not in JPEG land anymore
It’s web-only for now, and each video job costs around 8x more GPU time than an image job, roughly the same as an upscale.
That pricing, surprisingly, makes it 25x cheaper than anything else on the market so far!
As the company tests usage and demand, expect tweaks.
A "relax" mode for Pro subscribers is also in the works.
My static girl era is over, I want motion.