LTX-2.5

LTX-2.5 is an open-weights world model developed by Lightricks for generating synchronized video and audio from text, image and video inputs. It supports text-to-video, image-to-video and video-to-video modes with native audio generation and synchronization. Key capabilities include native multishot generation with character identity and style consistency across cuts, dynamic allocation of compute based on scene complexity, support for variable frame counts and resolutions up to 4K, and a custom Gemma 4 12B text encoder. A distilled version (8-step schedule) is available for faster inference.