VideoGen launches Model 1 for ten-minute videos and opens a $1 million credit pool
VideoGen's first in-house model plans and renders multi-shot videos with audio from one prompt. Access is limited to partners and credit applicants for now, and pricing is unpublished.
VideoGen, the Y Combinator-backed platform that has so far routed requests to other companies’ models, has announced its own: Model 1. Let’s Data Science reports that it launched on October 20 as a limited rollout to private partners.
According to VideoGen, Model 1 takes a prompt plus up to 16 reference images and writes the screenplay, shots and pacing itself, then generates video and audio together. The result is a finished multi-shot video up to ten minutes long. Most generators stop at 10 to 30 seconds per clip.
VideoGen claims Model 1 will cost less than half as much as comparable models. That rests on its own internal tests, and pricing and independent benchmarks are not out yet. Meanwhile VideoGen is handing out $1 million in free credits through the end of 2026. You apply from the Model 1 page, and it picks projects by scope and distribution plan.
For now, existing VideoGen plans (from $19 a month billed yearly) still run on third-party models. If you make long explainers or short films, apply for credits, but don’t buy a plan for Model 1 until it is generally available.
Source: Let's Data Science, Oct 20, 2026