LTX Video
Open source video generation model.
An open-source generative video model family from Lightricks for text-to-video and image-to-video generation, with local inference and fine-tuning workflows.
Pricing
Pay-as-you-go API billing; video generation is billed per second. Pricing varies by model, resolution, and operation. Official page: text-to-video and image-to-video LTX-2.5 Fast $0.09–$0.30/sec; LTX-2.5 Pro $0.12–$0.39/sec. Audio-to-video, retake, extend, HDR upscale, and reframe have separate listed rates.
Official LTX API pricing lists per-second rates that vary by model tier, resolution, and endpoint rather than a single price for the broadly named LTX Video product. Rates include LTX-2.5 Fast at $0.09–$0.30/sec and Pro at $0.12–$0.39/sec for text-to-video and image-to-video; additional endpoint and legacy-model rates are detailed on the page.
Model Intelligence
Recent stories
Amir Mushich used real product photos and the LTX-2.5 API to generate rotating product previews, then built the interface with Astra in Codex. Each preview links to the brand’s store.
Amir Mushich published a workflow for a web interface where clicks move between generated video states. It uses planned keyframes, matched forward and reverse clips, and a direct LTX API connection.
A creator used LTX-2.5 with uploaded audio, a start frame, and an action prompt to generate a continuous sequence across several shots. They report stronger prompting and continuity across wide, medium, and close views.
Reddit creators compared local AI video workflows for ComfyUI, including RTX workstation choices, inbetweening, talking-character sync, LTX2 generation, and Resolve edits. The setups point to local pipelines that combine generation, synchronization, and post-production tools.