Ltx v2 | Image to Video - AI Model

LTX‑2 is a comprehensive, open‑source AI video foundation model from Lightricks that natively generates synchronized audio and video. Built on a DiT (Denoising Diffusion Transformer) backbone, it offers granular creative control with multi‑keyframe conditioning, 3D camera logic, and LoRA fine‑tuning. It supports text‑to‑video, image‑to‑video, depth maps, and reference videos, delivering up to native 4K at 50 fps while running efficiently on consumer‑grade GPUs. Choose Fast, Pro, or Ultra modes to balance speed and fidelity, iterate quickly, and refine with advanced inputs for consistent style. LTX‑2 democratizes professional‑grade production by cutting compute costs and enabling 10‑second, high‑quality sequences.

Output Example

Used Prompt

A lone cyclist pedals fastly through a neon-lit city street at night, rain falling softly all around. The camera follows from behind with minimal motion, tracking the subtle movements of the bike as puddles splash beneath the tires. Reflections of neon signs shimmer across the wet pavement, fog rolling through the alleys. Occasional headlights pass by, casting streaks of light across the scene. The rain glistens in cinematic slow motion, droplets illuminated by pink and blue lights. Ultra-realistic 4K visuals, film grain, shallow depth of field, natural motion, moody ambient atmosphere, Blade Runner aesthetic.