LTXV-13B
13B open-weights video model — far faster than similar-sized rivals, trained on licensed footage.
About
LTXV-13B, released May 6, 2025, is Lightricks' bigger follow-up to the 2B LTX-Video: a 13-billion-parameter diffusion-transformer video model that keeps the family's faster-than-real-time ethos. Lightricks claims rendering up to 30x faster than similar-sized models, with multiscale rendering for finer detail control, VACE inference for video-to-video and reference-to-video editing, keyframe editing, and camera-motion control. The notable differentiator is the training data: licensed footage from Getty Images and Shutterstock, which sidesteps the copyright questions hanging over scraped-data video models. Weights are open under the LTXV Open Weights License, and Lightricks made commercial use free for enterprises under $10M in annual revenue. It runs on a single H100 or, with community Q8 kernels, on consumer hardware. Caveat: at 13B it is still heavier than the 2B original, and '30x faster' is Lightricks' own figure against unnamed rivals.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Lightricks-reported: up to 30x faster rendering than similar-sized models (provider-reported)
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (verified tier, metrics-v1). Evidence: Released May 6, 2025 per press (TechRound/TechStartups); 13B DiT video model from Lightricks.; Open weights (LTXV Open Weights License); trained on licensed Getty/Shutterstock data — notable copyright posture.; Free commercial license for enterprises under $10M revenue; runs on single H100 or consumer GPUs with community kernels.; Limitations: '30x faster' is Lightricks' own claim vs unnamed rivals..