VideoGen unveiled Model 1 on October 20, 2026 — its first AI video generation model, built to turn a text prompt or images into a finished, multi-shot video up to ten minutes long. Why it matters: most AI video tools still stop at short clips; a system that plans, generates and assembles a ten-minute piece in one workflow is aimed squarely at creators, filmmakers and advertisers who need complete videos, not fragments.
October 20, 2026
What happened?
San Francisco-based VideoGen, a Y Combinator-backed AI video company, announced Model 1 in a limited private-partner rollout starting October 20. General availability is expected to follow after additional safety testing. Alongside the model, the company opened applications for a $1 million pool of free AI credits for selected creative projects.
How does Model 1 work?
Users start with a written brief and can add up to 16 images as visual references, along with a target duration and aspect ratio. From there, the system handles creative planning, audio and video generation, and automated assembly in a single workflow. Its planning layer develops the screenplay, dialogue, shot list and pacing, and carries context across scenes so characters, locations, visuals, audio and story stay consistent from shot to shot. Longer videos are built by generating individual shots and then joining the selected clips.
Under the hood, the company says the model reasons about picture and sound together and keeps reusable representations of recurring characters, objects and locations. The pipeline also checks its own output for problems such as incomplete dialogue or visual inconsistencies and can retry selected renders before assembling the final cut.
What about cost and safety?
VideoGen says that, based on its own internal validation, it expects Model 1 to cost more than 50% less than comparable models. That figure needs context: final pricing has not been published, and the company says official benchmarks and third-party validation are still underway. Treat the cost claim as a company estimate for now, not a tested result.
On safety, the release says Model 1 includes guardrails designed to block adult and illegal content, and that generated outputs carry invisible watermarks identifying them as synthetic.
How does the $1 million credit program work?
Applications opened October 20, and VideoGen plans to distribute the credits through the end of 2026. Individual allocations will depend on each project’s scope and scale, and selection will consider the creative work itself and whether there is a viable distribution plan. Professional creators, filmmakers and studios are eligible, and so are students and amateurs.
Why it matters
For Tested Intelligence readers, the practical question is not whether AI can make a striking ten-second clip — it already can. The question is whether it can hold a story together for ten minutes without the characters, locations and audio drifting apart. Model 1 is a direct attempt at that harder problem, and its credit program is a low-cost way for creators to test it on a real project. The honest caveats are the ones the company itself flags: pricing is not final, benchmarks are not independently validated yet, and general availability still depends on further safety testing. Everything AI, tested — which here means waiting for hands-on results before calling the cost and quality claims proven.
Frequently asked questions
How long a video can VideoGen Model 1 make?
Up to ten minutes, assembled as a finished, multi-shot video from a text prompt and optional images.
When can anyone use Model 1?
A limited private-partner rollout began October 20, 2026. General availability is expected after additional safety testing; no exact date was given.
Who can apply for the $1 million in free credits?
Creators, filmmakers and studios, including students and amateurs. Credits will be distributed through the end of 2026 based on project scope and distribution plans.
Sources: GlobeNewswire (VideoGen, Inc. release, October 20, 2026)