Midjourney has released its first video generation model, V1.

Startup Midjourney, known for its AI-based image generation tools, officially launched its first video model generator — V1 — on June 18. The new product operates on the "image to video" principle: users upload an image, and the system creates four five-second clips.

Like other company products, V1 is available exclusively through Discord and currently works only in the web version. Midjourney plans to further develop its generative video direction — from current animations to full-fledged realistic real-time simulations. In the future, the company intends to launch models for generating 3D objects and interactive scenes.

With the release of V1, Midjourney enters competition with industry leaders: Sora from OpenAI, Gen-4 from Runway, Firefly from Adobe, and Veo 3 from Google. However, unlike competitors focused on commercial use, Midjourney emphasizes creativity and visual style.

The V1 model allows users to control video parameters: set random or text-described animations, adjust the degree of camera and object movement, and extend clips up to 21 seconds. By default, clips last five seconds.

The cost of generating video is eight times higher than generating images. Subscribers on the basic $10/month plan will be able to test the model, but their limit will quickly run out. Users on the Pro — $60 and Mega — $120 plans will have unlimited access to video generation in Relax mode. Over the next month, the company plans to revise pricing for video models.

Source: TechCrunch