Midjourney has launched its first AI video generation model, V1, marking its entry into a growing field already populated by major players like OpenAI’s Sora, Runway’s Gen 4, Adobe’s Firefly, and Google’s Veo 3. Known for its creative-focused AI image models, Midjourney is now expanding into video, aiming to go beyond typical commercial uses like advertising and B-roll.
CEO David Holz shared in a blog post that the company envisions eventually creating AI tools capable of powering real-time open-world simulations. Along with video generation, future plans include tools for 3D rendering and real-time AI models.
The release of V1 comes just a week after Midjourney was hit with a lawsuit from Disney and Universal, who claim its AI-generated images feature copyrighted characters like Darth Vader and Homer Simpson.
Hollywood studios have raised concerns over the use of copyrighted material in training AI models, especially as these tools grow in capability and popularity. Despite branding itself as a creativity-first company, Midjourney remains under scrutiny alongside other AI content platforms.
At launch, the V1 model includes customizable features such as automatic or manual animation controls, as well as options to adjust the amount of movement within videos. The default video length is five seconds, but users can extend clips by four seconds up to four times, reaching a maximum of 21 seconds. True to Midjourney’s style, the early V1 videos appear surreal and artistic rather than ultra-realistic.
Access to V1 starts at $10 per month with the Basic plan, but subscribers to higher-tier plans — $60 Pro and $120 Mega — get unlimited video generation in “Relax” mode. However, video generation costs eight times more than image generation, meaning users may use up their allowances much faster. The company plans to revisit its pricing strategy over the next month.
Though V1 is still new, it has received a warm response for its unique visual style. How it compares to more established AI video models remains to be seen.
