Lightricks Extends AI Video Length with LTXV Update

Lightricks has announced a major update to its LTXV model, breaking the 60-second barrier for AI-generated video length, announced in a press release. This advancement positions Lightricks as the first company to offer real-time, long-form AI video creation at scale, surpassing the previous industry standard of eight seconds.

The updated LTXV model features an autoregressive streaming architecture that allows for continuous video generation in real time. This enables developers and creators to produce longer, coherent narratives rather than isolated clips. The new capability is available to developers, academia, and AI video enthusiasts, and will also power commercial products like LTX Studio, Lightricks' flagship creative development platform.

The LTXV update introduces an improved video engine that streams video live to viewers as it is rendered, with the first second of content returned almost instantly. This allows for interactive, continuous storytelling with full control over scene development, paving the way for new generative storytelling applications in advertising, gaming, live events, and education.

Lightricks' LTXV model supports both 13B and mobile-friendly 2B parameter models, and is compatible with IC-LoRA motion and style integrations. The system runs efficiently on a single H100 or consumer-grade GPU, delivering artifact-free 30-second clips, making it accessible for a wide range of users.