Lightricks unveils 13B LTX video model for HQ AI video generation
Lightricks, a pioneer in AI-powered creative technology, has officially launched its most advanced generative video model to date — the LTX Video 13-billion-parameter system, or LTXV-13B. This milestone signals a major upgrade in the company’s flagship product line, promising superior quality, unmatched rendering speed, and broader accessibility across its ecosystem. The LTXV-13B model is now available within LTX Studio, the company’s core storytelling platform, and is being integrated across Lightricks’ suite of tools.
LTXV-13B marks a significant technological leap forward for Lightricks, leveraging a 13-billion-parameter architecture designed to unlock new levels of creative precision, video fidelity, and performance. The model merges cutting-edge academic research with recent innovations in open-source communities, integrating advanced mechanisms like unsampling controls and spatial-temporal guidance for enhanced video editing workflows. These enhancements have positioned the model to compete aggressively in the fast-evolving market for generative video AI.
What separates LTXV-13B from its competitors is its ability to deliver remarkably high-quality video output while sustaining fast rendering speeds. A major factor behind this performance is Lightricks’ new “multiscale rendering” architecture — a rendering methodology inspired by the creative workflows of visual artists and filmmakers.
Multiscale rendering mimics the way a traditional artist might sketch a scene in broad strokes before layering in color, depth, and finer details. In technical terms, this means the model initially generates a fast, physically accurate base layer of the scene and then applies high-resolution detail passes in a second stage. This iterative process not only enhances visual coherence and resolution but also slashes latency to a fraction of what similarly sized models require. According to Lightricks, render times using LTXV-13B can be up to 30 times faster than those of competing systems of comparable size.
The company has also engineered the model with real-world accessibility in mind. LTXV-13B is optimized to run efficiently on a range of consumer-grade GPUs, a move that democratizes AI video generation by enabling creators to work on personal devices without needing cloud-scale infrastructure or professional-grade hardware.
“The introduction of our 13B parameter LTX Video model marks a pivotal moment in AI video generation with the ability to generate fast, high-quality videos on consumer GPUs,” said Zeev Farbman, co-founder and CEO of Lightricks. “Our users can now create content with more consistency, better quality, and tighter control. This new version of LTX Video runs on consumer hardware, while staying true to what makes all our products different — speed, creativity, and usability.”
To further expand its model’s capabilities and dataset integrity, Lightricks has entered key strategic partnerships with top media providers. A notable collaboration was established with Getty Images, following a similar agreement with Shutterstock announced in December 2024. These alliances give Lightricks access to a vast archive of premium visual content for model training, supporting its ambition to build tools that are not only powerful and visually rich but also ethically trained and safe for commercial use.
The LTXV-13B platform now features broader functionality that allows for deeper creative control, supporting high-level production features such as:
- Keyframe editing
- Camera motion control
- Character and scene-level motion adjustment
- Multi-shot sequencing and editing
Lightricks is also extending free licensing of the LTXV-13B model to enterprises with annual revenues under $10 million, opening the doors for smaller creative teams and emerging startups to integrate advanced AI video tools without the financial barrier of enterprise licensing. In parallel, Lightricks has committed to open-sourcing its LTXV models, now publicly available on Hugging Face and GitHub under the names LTX-Video.
“By consistently refining our models and working with the open community, we’ve built an AI system that generates physically natural movement while preserving artistic control,” added Yoav HaCohen, Director of LTX Video at Lightricks.
Since its initial launch in November 2024, LTX Video has seen continuous refinement, supported by contributions from academic researchers and the broader open-source developer community. This collaboration has led to steady improvements in motion smoothness, temporal consistency, and overall creative flexibility.
Some of the notable open-source innovations incorporated in LTXV-13B include:
- Vace Model Inference – advanced video generation and editing tools, including reference-to-video (R2V). Details on GitHub
- Upsampling Controls for Video Editing – Tools that reverse noise and refine frame granularity. Details on GitHub.
- Kernel Optimization – UEfficient Q8 kernel usage allows performance scaling on lower-resource devices. Details on GitHub and HuggingFace.
With the release of the LTXV-13B model, Lightricks is doubling down on its mission to redefine what’s possible in generative AI video. The company’s strategy—built on performance, accessibility, and collaboration—positions it at the forefront of a new era in AI-powered creative media. For more on Lightricks’ open-source initiatives and access to their models, visit www.lightricks.com.