Midjourney Announces First Video Generation AI Model 'V1,' Capable of Generating Videos of up to 21 Seconds from a Single Image

Midjourney, a developer of image generation AI, has announced the release of its video generation AI model ' V1 ', available via Discord. V1 can generate four 5-second videos based on images simply by using images uploaded by you or images generated by other models in Midjourney.

Introducing Our V1 Video Model
https://www.midjourney.com/updates/introducing-our-v1-video-model

Midjourney launches its first AI video generation model, V1 | TechCrunch
https://techcrunch.com/2025/06/18/midjourney-launches-its-first-ai-video-generation-model-v1/

Below is a movie that actually edits and connects the videos generated by V1. The usage fee is $10 (about 1,450 yen) per month.

Introducing our V1 Video Model. It's fun, easy, and beautiful. Available at 10$/month, it's the first video model for *everyone* and it's available now. pic.twitter.com/iBm0KAN8uy

— Midjourney (@midjourney) June 18, 2025

The following video was also generated by V1. V1 can generate four 5-second videos at a time, and the generated video can be extended up to 4 times for 4 seconds. In other words, V1 can generate videos up to 21 seconds long.

midjourney introduces video generation and it's surpassing all my expectations. pic.twitter.com/lWqakbSTVV

— Phi Hoang (@apostraphi) June 18, 2025

The animation generated from the illustration in V1 looks like this.

Midjourney V1 video model i2v test pic.twitter.com/J6GVJkH55O

— naginagi???? (@nagin2ko) June 18, 2025

On X (formerly Twitter), some users were comparing videos generated by V1 with those of a competing video generation AI, Runway .

Finally Midjourney v1 video model released✨
Have you tried it????

I think Midjourney looks better when you focus on the people. I'll compare it to Runway, but Runway seems to capture the whole picture.
For comparison, the prompts are identical. pic.twitter.com/lOAVnXc8M0

— sami (@syamiyo) June 18, 2025

V1 is a model that can generate 'Image-to-Video', and you can generate images as usual in Midjourney and add movement to the images by pressing the 'Animate' button.

In addition, if you set the animation setting to 'auto', a motion prompt that specifies the movement of the video will be automatically generated, and if you set it to 'manual', you can instruct the system to move objects and develop the scene, just for videos that just move objects. Other settings available include 'low motion', where the camera is almost still and the subject moves slowly, and 'high motion', where the subject and camera move a lot.

In addition, V1 can animate not only images generated by Midjourney, but also images uploaded from outside Midjourney. Simply drag an image onto the prompt bar to mark it as the 'start frame,' enter a motion prompt, specify how you want it to move, and a video will be generated.

It costs about eight times as much to generate a video as it does to generate an image. Nevertheless, Midjourney claims that the cost of one second of video is equivalent to the cost of one image. This is an astonishing figure, more than 25 times cheaper than what the market has offered so far.

While many companies are focused on developing video generation AI for commercial use, Midjourney has been developing its own image generation AI for creative types. However, with the release of V1, Midjourney will compete with other video generation AI models such as OpenAI's Sora , Adobe's Firefly , Google's Veo 3 , and Runway Gen 4 .

Midjourney CEO David Holtz has big goals for its video generation AI beyond producing B-roll for Hollywood movies or commercials for the advertising industry, and says V1 is a step toward its ultimate goal of 'building an AI model capable of real-time open-world simulation.'