On June 18, 2025, Midjourney — the company that built its name on image generation — released V1, its first AI video model. For a brand that turned aesthetic consistency into a loyal community, this was the missing product line, and the formal answer to Google’s Veo 3, OpenAI’s Sora, and Runway’s Gen 4, all of which had been shipping for months.
V1 is an image-to-video system rather than a text-to-video generator: you feed it a picture, either generated by Midjourney’s own image models or uploaded from elsewhere, and it animates the still frame. It is not cheap, but as Engadget relayed the company’s framing, this was “the first video model for everyone” — aimed at individual creators rather than Hollywood production pipelines.
How V1 works
The mechanics: each generation produces four five-second clips. Each clip can be extended by four seconds at a time, up to four extensions, for a maximum of roughly 21 seconds. Two animation modes cover intent — automatic animation lets the model decide how the scene moves, while manual mode takes a text description of the camera work and motion you want and follows it.
A low-motion and a high-motion setting control how much the camera and subject move. TechCrunch’s early tests described the output as somewhat otherworldly rather than cinematic realism, with reactions from the first wave of users broadly positive. The model was available at launch on Midjourney’s web platform and in Discord, plugging directly into its existing image workflow. For users already inside that interface, the learning curve was close to zero: the video tools sat next to the image tools, ran on the same subscriptions, and kept the community in one place — a quiet but real advantage against rivals that required separate apps or APIs.
Pricing and usage limits
Video generations cost roughly eight times as much as image generations — the headline constraint on V1 usage. Entry is the $10-per-month Basic plan; Pro subscribers at $60 and Mega subscribers at $120 per month get unlimited video generations in the slower Relax mode, with the effective ceiling set entirely by which tier you pay for.
Midjourney said the pricing would be reassessed within a month — an acknowledgment that video compute costs were still being figured out, and a hedge against mispricing a new compute-heavy product. The contrast with rivals’ billing was stark: Veo 3 sat behind Google’s premium subscriptions and API pricing, Runway billed by credits, and Midjourney kept video inside its own subscription wall, leaning on its existing image community to absorb the cost. Light users paid by usage, heavy users went unlimited on a plan — the same playbook that had built its image business.
The competitive picture: late but different
V1 arrived as a chaser, and that is worth stating plainly. Sora, Runway’s Gen 4, Adobe’s Firefly, and Google’s Veo 3 had been on the market for months or longer, with Veo 3 in particular dominating the conversation thanks to native audio and realism. Midjourney’s differentiation was deliberate: no commercial B-roll production lines, just stylized video aimed at creators’ visual taste.
TechCrunch quoted CEO David Holz framing video as a step toward models “capable of real-time open-world simulations,” with 3D rendering and real-time models planned next. In that framing, V1 mattered less as a video-market competitor and more as the data and engineering stepping stone toward real-time simulation — a long-term declaration worth more than V1’s spec sheet.
Copyright shadow and the roadmap
The timing was awkward: a week earlier, Disney and Universal had sued Midjourney for copyright infringement, alleging its outputs depicted protected characters including Homer Simpson and Darth Vader. Alongside the video launch, Midjourney asked users to “please use these technologies responsibly” — in the middle of that lawsuit, the request read as more than a formality. Shipping anyway carried a statement of its own: the product roadmap would not pause for litigation, and every V1 showcase doubled as proof the launch train kept moving.
From 2026, V1 reads as the start of Midjourney’s turn from an image company into a multimodal creation platform. It was not the most capable video model of its moment, but the combination of a moderate entry price and unlimited Relax-mode generation put it within reach of the existing subscriber base, and made it a useful sample of how creator markets actually adopted video generation.
Sources
- Midjourney launches its first AI video generation model, V1 - TechCrunch
- Midjourney adds AI video generation - Engadget
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.