Lightricks has made a major leap in AI-powered video generation with the release of LTX-2, its latest open-source foundation model. The company claims the model can create six-second, high-definition clips in just five seconds — faster than playback speed — while also supporting native sound generation and 4K rendering.
This marks a significant milestone for creators seeking rapid, high-quality, and cost-efficient video production without relying on expensive studio hardware.
Accelerated Creativity for Everyone
The LTX-2 model stands out for more than just its speed. It merges video and audio synthesis in one workflow, allowing users to generate synchronized visuals, dialogue, and ambient sound effects simultaneously. Previously, creators had to craft video and audio separately, then manually sync the two — a process that often slowed down creative iteration.
Now, Lightricks’ diffusion-based system enables near-instant previews. Creators can quickly test ideas, tweak prompts, and refine scenes, bringing AI-driven storytelling much closer to real-time production.
Even more impressive: users with consumer-grade GPUs can run the model locally, drastically cutting compute costs and making professional-level video generation accessible to a broader audience.
Diffusion Models Come of Age
At its core, LTX-2 is a diffusion model — a type of AI that starts with noisy data and refines it step by step into a coherent image or video. Lightricks has optimized this process for speed and stability, proving that diffusion models can now meet the demands of professional content creation.
Zeev Farbman, Lightricks’ co-founder and CEO, describes LTX-2 as “the most complete creative AI engine we’ve ever built,” highlighting its combination of synchronized audio-video generation, 4K fidelity, flexible workflows, and radical processing efficiency.
He adds, “This isn’t vaporware or a research demo — it’s a real breakthrough in video generation.”
Professional Quality Meets Open Access
For professionals, the LTX Studio platform provides advanced creative control over rendering parameters — including pose, depth, and video-to-video generation. While high-end 4K output at 48 frames per second may take slightly longer to process, the result is cinematic quality at a fraction of traditional production costs.
The company plans to release further customization options soon, giving users granular control over the creative pipeline and model behavior.
A Legacy of AI Innovation
LTX-2 builds on Lightricks’ earlier LTXV series, which introduced long-form video generation up to 60 seconds, real-time prompt updates, and multi-scale rendering. Those advancements allowed users to progressively enhance scenes — similar to how professional animators layer visual detail during post-production.
The 13B model from that series was trained on licensed data from Getty Images and Shutterstock, a key ethical move that reduces copyright concerns plaguing other AI video generators.
The company has also experimented with model distillation, allowing faster generation in as few as four steps and supporting LoRAs for project-specific fine-tuning.
Pricing, Access, and Open-Source Transparency
Lightricks is taking a hybrid approach: LTX-2 is open-source, with its full model weights, datasets, and tools set to appear on GitHub. This makes it one of the few genuinely transparent alternatives in an increasingly closed-off AI video market.
For those who prefer hosted access, the company offers multiple paid tiers via API:
- HD generation starts at $0.04 per second
- Pro tier costs $0.07 per second for balanced quality and speed
- Ultra delivers 4K at 48fps with full audio for $0.12 per second
Lightricks says these prices make LTX-2 up to 50% more affordable than comparable AI video tools — a major factor for studios and independent creators managing long projects.
The open-source release arrives next month, but users can already test LTX-2 today through LTX Studio or the API.
The Bigger Picture
With LTX-2, Lightricks isn’t just releasing another model — it’s redefining what’s possible for AI-assisted video creation. By merging open-source accessibility, 4K resolution, synchronized audio, and consumer-friendly performance, the company is helping AI video generation cross from experimental novelty to true production-grade utility.
As AI tools continue to democratize creative workflows, Lightricks’ latest model positions itself as both a challenger to industry giants like Google’s Veo and a catalyst for the next generation of independent creators.
