VideoGen, a San Francisco-based, Y Combinator-backed AI video company, has unveiled VideoGen Model 1, its first AI video generation model. The model turns text prompts and images into finished, multi-shot videos up to ten minutes long — a major jump over the short clips most AI video tools produce today.
Model 1 brings creative planning, audio and video generation, and automated assembly into one workflow. Users start with a written brief and optional visual references (up to 16 images), and an AI “thinking harness” develops the screenplay, dialogue, shots, and pacing — carrying context across scenes to keep characters, locations, and narrative consistent. The architecture uses joint audio-video attention to model visuals and sound together, and the system checks its own work for incomplete dialogue and visual inconsistencies, retrying renders within a set budget.
Based on internal validation, VideoGen expects Model 1 to cost more than 50 percent less than comparable models, with official benchmarks and third-party validation still underway. The company also opened applications for a $1 million pool of free AI credits for selected creators, filmmakers, and studios — students and amateurs included — with credits distributed through the end of 2026.
The rollout began October 20 as a limited private-partner release, with general availability expected after additional safety testing. Model 1 includes guardrails against adult and illegal content, and generated outputs carry invisible watermarks identifying them as synthetic. The release puts VideoGen alongside heavyweights like Sora, Pika, Runway, and InVideo in the race for long-form AI video.
Source: GlobeNewswire
