LTX Video 2.0 Full Quality

• 🆕 Lightricks' open-source DiT model generating synchronized video and audio.
• 👁️ Image-to-video: animates a source image from a motion prompt.
• 📏 Targets native 4K fidelity at 50 fps, up to 10-second clips.
• 🎯 Part of a multi-mode LTX-2 lineup with fast and distilled variants.
• ⚡ Designed to run on consumer-grade GPUs for local execution.
• 🔧 Open weights plus training and inference code released.
• 🌐 Integrates with Diffusers and ComfyUI workflows.

Lightricks

Lightricks is a creative technology company founded in January 2013, widely recognized for its popular selfie-editing app Facetune. Originally focused on mobile video and image editing tools, the company has expanded significantly into generative AI,…

LTX Video 2.0 Full Quality is the high-fidelity, image-to-video member of Lightricks' LTX-2 family — a DiT-based audio-video foundation model that generates synchronized video and sound within a single system. Given a source image and a prompt describing motion, camera movement, and audio, it produces a clip. Lightricks released open weights alongside training code and inference pipelines.

Relative to the company's earlier LTX-Video (LTXV), which was a real-time DiT video generator without sound, LTX-2 adds natively synchronized audio, targets 4K output at 50 fps, and supports roughly 10-second sequences while remaining runnable on consumer GPUs, per Lightricks' model card. The lineup spans multiple performance modes, so this Full Quality variant pairs with LTX Video 2.0 Fast for quicker previews and LTX Video 2.0 Full Quality (text-to-video) for prompt-only generation.

Within the same image-to-video family, Lightricks later shipped LTX Video 2.0 19B and, more recently, LTX Video 2.3 Full Quality, which the catalog lists as the newest full-quality iteration. Lightricks frames the family as evolving through real-world use, with community-contributed LoRAs and ComfyUI workflows extending its capabilities.

Note: detailed independent benchmark figures for this specific variant were not available from primary or top-evaluator sources, so this description stays factual.