LTX
Category: Generative Media (Image / Video)
Open world models company spun out of Lightricks, building open-weights foundation models for video generation and physical world simulation. LTX was founded in 2024. The company is led by Zeev Farbman. Based in Jerusalem, Israel. Team size: 101-500. Total funding raised: $150M. Latest round: Corporate Round. Key investors include Lightricks (parent company).
- Founded
- 2024
- Headquarters
- Jerusalem, Israel
- Team size
- 101-500
- Total funding
- $150M
Value proposition
Production-ready, open-weights multimodal video and world models that enterprises can run, fine-tune, and deploy on their own hardware — delivering native 4K, synchronized audio-video, up to 50fps at roughly one-eighth the cost and one-tenth the render time of comparable closed models.
Products and solutions
LTX-2.5 (latest open-weights video/world model, generates 10-sec clips from images in ~6.8s on Nvidia superchips), LTX-2, LTX-2.3, LTXV models, LTX Studio (AI filmmaking platform), LTX API, LTX Desktop, LTX Trainer, LTX Playground.
Unique value
Open-weights world models with production-grade output (native 4K, synced audio-video, up to 50fps) at a fraction of the cost and render time of closed competitors, with full enterprise control over IP and local/private-cloud deployment.
Target customer
Enterprises, film/TV studios, VFX houses, advertising agencies, brand studios, gaming companies, and robotics/physical AI startups.
Industries served
Media & Entertainment, Film/VFX, Marketing & Advertising, Gaming & Interactive, Robotics & Physical AI, Real-time AI.
Technology advantage
Proprietary diffusion architecture, VAE design, and real-time inference stack (built by CTO Yaron Inger's team); most-downloaded open world model line on Hugging Face with 33M+ downloads; native ComfyUI integration; efficient inference on Nvidia superchips.
How they differentiate
Unlike closed competitors (Sora, Veo, Runway), LTX is fully open-weights, allowing enterprises to run, fine-tune, and extend models on their own hardware with full IP control, while delivering production-grade output at dramatically lower cost and render time.
Main competitors
Runway, Luma AI, Kling AI (Kuaishou), Google Veo, OpenAI Sora, ByteDance Seedance, Pika, Hailuo AI.
Key partnerships
Nvidia (superchip inference), ComfyUI (native integration), Hugging Face (open model distribution), Google (official partner), Lightricks (parent company).
Notable customers
L'Oréal–McCann Paris, eToro, ElevenLabs, Clutch, Hear Well Be Well, CYLNDR Studios, VW Trucks, Taika Waititi (filmmaker).
Major milestones
Founded 2024 at Lightricks, LTX-2 launched Jan 2026 (top open model on Hugging Face), LTX-2.5 released (10-sec video from image in 6.8s on Nvidia superchips), officially spun off from Lightricks as independent open world models company (June/July 2026).
Growth metrics
33M+ Hugging Face downloads across LTX model family (claimed at LTX-2.5 launch Aug 2026); was 22M+ after LTX-2 (Jan 2026) with 3M in first month; most-used open world model line; 135K+ producers/creative teams use LTX Studio.
Market positioning
Most-downloaded open world model line on Hugging Face (33M+ downloads); LTX-2 Pro previously ranked #3 Image-to-Video on Artificial Analysis (Nov 2025; rankings dynamic—LTX-2.5 now mid-pack overall); positioned as the leading open alternative to closed video/world model leaders.
Geographic focus
Global, with enterprise focus across US, Europe, and Asia; headquartered in Jerusalem, Israel.
About Zeev Farbman
Co-founder & CEO of Lightricks (led to ~500-person team and $1.8B valuation); PhD in Computer Science from Hebrew University of Jerusalem (research on image optimization). Founded LTX in 2024.
Latest news about LTX
More Generative Media (Image / Video) companies
Official website: https://ltx.io
