Lightricks

Introducing Retake in LTX Built to edit a shot after it's rendered. Retake brings true directorial control to AI filmmaking, for the first time. To show what this looks like in practice, we recreated one of our friends at monday.com's YouTube ads and changed just one element: the opening line (The Hook) Everything else stayed exactly the same. Same footage. Same pace. Same edit. Same actor. Same voice. No reshoots. No new production cycle. No waiting for the perfect take. Just a single sync and suddenly you can A/B test infinite versions from the footage you already have. This is creative agility in action, now integrated directly into LTX, and through the API.

A little stache goes a long way … We’re bringing awareness to Men’s Health Awareness Month AKA #Movember using the power of LTX-2 (and a killer AI-generated mustache). Huge shoutout to our very own Tom Factor, Senior Director of LTX Creative Marketing, for sliding this masterpiece into our Slack DMs … we couldn’t help but share it with you 😉

Live now at GenML 2025! Our team is showing what next-gen video models can really do. Ofir Bibi, VP Research at Lightricks, is on stage unpacking LTX-2, our video foundation model - from architecture to training to what makes it stand out. Naomi Ken Korem (Vinokurov), Senior Researcher at Lightricks, is leading a hands-on workshop on fine-tuning text-to-video models with LoRAs, helping creators and developers take full control of AI video generation. If you’re at #GenML2025, come say hi, we’d love to connect, talk research, and swap ideas on where AI video is headed next.

Yesterday, we had the privilege of hosting the launch of The AI Collective - Jerusalem, bringing together over 100 brilliant minds from across the city’s growing AI ecosystem. We see it as part of our mission not only to advance technology, but to help build the communities that make meaningful innovation possible. During the event, Naomi Ken Korem (Vinokurov) from our research team shared how Lightricks is pushing the boundaries of AI video generation, combining cutting-edge research with real creative impact, and showing how open collaboration drives progress across the ecosystem. A special thank you to Mouhammad J. and Max Friedman for leading this initiative and helping put Jerusalem’s AI scene on the global map.

How does a five-person startup that redefined visual creativity grow into a leader in AI video generation? In the latest podcast episode of "Founders and Builders podcast", Yaron Inger, our Co-founder & CTO, shares the journey - from the early days of Lightricks to developing LTX-2, our new open-source AI video foundation model. Yaron dives into what it takes to build technology that feels creative, how open-source collaboration is shaping the next wave of AI video, and why accessibility and delight remain core to everything we build: from apps to production-grade AI models. If you’re interested in the intersection of AI, creativity, and product innovation, this one’s worth a listen. 🎧 Link to the podcast in the comments. 

“AI video isn’t just about what models can generate - it’s about who gets to create.” In the latest episode of Hidden Layers, Naomi Ken Korem (Vinokurov), Senior Researcher at Lightricks, takes us inside the fast-evolving world of AI video generation, and how open-source models are transforming creative workflows. We discuss what it really takes to train foundation models for video, their advantages over competing approaches, and where this technology is headed next. Naomi also shares how our research journey at Lightricks evolved from breakthrough consumer apps to AI models built for real production, combining high fidelity, speed, and accessibility in one workflow. It’s a conversation about AI as a true co-creator, and the shift from closed systems to open tools that let anyone, from studios to indie teams, build, test, and create at scale. Link to the full episode in the first comment.

What an incredible night we had at GenJam: New York Festivals x Machine Cinema, presented by LTX and 4As, and hosted at McCann New York. Seeing creativity and technology collide, while pushing the boundaries of what’s possible with generative AI, is what we are all about. In just a few hours, teams transformed imagination into motion, showing how AI can inspire new forms of storytelling and expression. At Lightricks, we believe AI is here to empower creators, not replace them, and GenJam was the perfect showcase of that vision. A huge thank you to everyone who joined, participated, and brought their creative energy to the room. Thanks to our team on the ground who took part. Check out the highlights from GenJam below. Shani Mandel-Laufer Tian Pei Joshua Pasternak Ann Streichman Adam Palmer Alaysia Brown

LTX-2 just hit #3 globally on Artificial Analysis’ Video Arena - the independent benchmark comparing the world’s leading AI video models. It’s an incredible milestone for the team at Lightricks, placing LTX-2 alongside the biggest names in AI. Our LTX-2 ranks #3 in Image-to-Video, and LTX-2 sits at #4 in Text-to-Video, or #3 when grouped by provider. And these rankings are based on video-only evaluations, before accounting for LTX-2’s multimodal capabilities (audio + video), currently shared by only a few providers: Google, Alibaba, OpenAI, and Lightricks. Built for real production, LTX-2 unifies synchronized audio and video generation in one pipeline for cinematic timing and storytelling. It delivers native 4K output at up to 50 fps, multi-keyframe conditioning, camera logic, and LoRA fine-tuning, all optimized for efficient, scalable performance on affordable hardware. This is a powerful validation of our belief that frontier-level foundation models can be fast, efficient, and open - and we’re just getting started. Check out the Artificial Analysis ranking in the first comment below.

Our booth at this year’s Reversim Summit drew hundreds of developers who got hands-on experience with LTX-2, our in-house developed #AI video model. It was inspiring to see the excitement around how we build and train our models from the ground up. And yes, apologies to those who had to wait in line to receive our swag, we guess that’s what happens when your booth is the busiest one at the conference. 💫 Thank you to everyone who stopped by, asked great questions, and engaged with our team. See you next year! 📸: Victor Levy & Roey Shor