VideoGen Unveils Model 1, Its First AI Video Generation Model: Up to 10-Minute Films From a Single Prompt

What happened

According to VideoGen’s official October 20 press release (via GlobeNewswire), Model 1 supports text-to-video and image-to-video with up to 16 reference images, and users can specify the target duration and aspect ratio. The model runs joint audio-video attention in latent space, keeping characters, locations, visuals, audio and narrative consistent across independently generated shots; outputs carry invisible watermarks identifying their synthetic origins, and safety guardrails are designed to block adult and illegal content.

On rollout and pricing: a limited private-partner rollout began October 20, with general availability to follow after additional safety testing; final pricing, official benchmarks and third-party validation are still underway. The separate $1 million free-AI-credit pool is open to creators, filmmakers, studios, students and amateurs — applications opened October 20, with allocations based on project scope and credits distributed through the end of 2026.

Key facts

1. Length:a single prompt or image yields finished multi-shot films up to 10 minutes long; users can specify target duration and aspect ratio.
2. Inputs:text-to-video plus image-to-video with up to 16 reference images; the AI “thinking harness” plans screenplay, dialogue, shots and pacing, keeping characters, locations, visuals, audio and narrative consistent across scenes.
3. Architecture:joint audio-video attention in latent space; in an internal experiment an eight-step adapter cut average clip-render time by 58.2% versus a 28-step baseline (experimental result, not guaranteed); a technical whitepaper is forthcoming.
4. Cost:the company expects costs more than 50% below comparable models (internal validation); final pricing, official benchmarks and third-party validation are underway.
5. Safety:guardrails designed to block adult and illegal content; outputs carry invisible watermarks identifying their synthetic origins.
6. Availability:limited private-partner rollout began October 20; general availability follows after additional safety testing.
7. Credits:a $1 million pool of free AI credits for creators, filmmakers, studios, students and amateurs; applications opened October 20, allocations by project scope, distributed through end of 2026.