Lightricks, the company behind AI-driven tools Facetune and LTX Studio, unveiled an open-source generative AI video model, LTX Video 0.9 (LTXV), that promises to revolutionize video creation. The Israel-based company has made LTXV freely available to researchers and developers with an aim of advancing its AI video technology.

The model boasts an unprecedented ability to create five-second video clips in just four seconds—a processing speed faster than playback time. Designed with accessibility and efficiency in mind, LTXV is compatible with widely available GPUs, bringing cutting-edge video generation capabilities to smaller studios and independent creators.

LTXV is a DiT-based video generation model that makes high quality video in real time. The model can generation 24 FPS videos at 768x512 resolution, and it is trained on a large-scale dataset. Built with a Diffusion Transformer architecture, it ensures smooth and coherent transitions between frames, eliminating common artifacts like object morphing.

“We built Lightricks to bridge the gap between imagination and creation,” said Zeev Farbman, Co-founder and CEO of Lightricks. “With LTXV, we’re providing an open-source tool that empowers researchers and developers to shape the future of AI video creation.”

Lightricks has released LTXV on GitHub and Hugging Face, built using PyTorch XLA to ensure compatibility across GPU and TPU systems. This open-source approach mirrors the company’s earlier release of Long AnimateDiff, am animation framework.

The integration of LTXV into LTX Studio significantly enhances the platform, enabling creators to produce longer, high-quality videos at unprecedented speeds. By eliminating technical hurdles, LTXV allows users to focus on creativity and storytelling, paving the way for new possibilities in content creation.

“LTXV represents a new era of AI-generated video,” said Yaron Inger, CTO of Lightricks. “Its unparalleled speed and precision open the door for interactive experiences in gaming, shopping, and social engagement. We’re excited to see how the global community builds upon this model.”

LTXV delivers results comparable to larger, more resource-intensive models. Some example clips created using the model are listed in Github, and Lightricks plans future updates to enhance motion consistency, support higher resolutions, and expand customization options.