Lightricks has unveiled LTX-2, an open-source AI video model that dramatically speeds up content creation. The model can generate six-second stylized videos in just five seconds, with optional 4K resolution at 48fps and fully synchronized audio. LTX-2 uses diffusion technology to iteratively refine video and audio, enabling near-instant previews of results.
It runs efficiently on consumer GPUs, reducing costs, and offers flexible creative controls such as depth, pose, and video-to-video generation. LTX-2 will be available via LTX Studio and API, with an open-source version planned for GitHub, combining speed, high quality, and cost-efficiency for professional video workflows.