LTX-Video

> [!IMPORTANT]
> **LTX-2.5 is the latest LTX model and is recommended for all new projects.**
>
> Released on August 11, 2026, LTX-2.5 supersedes both LTX-Video (v0.9.x) and LTX-2. It brings native multi-shot generation, synchronized audio and video, stronger prompt adherence, and better local performance.
>
> | Resource | Link |
> |---|---|
> | LTX-2.5 code and weights | https://github.com/Lightricks/LTX-2 |
> | LTX-2.5 model overview | https://ltx.io/model/ltx-2-5 |
> | LTX-2.5 documentation | https://docs.ltx.io/models/ltx-2-5 |
> | LTX-Video v0.9.x models | This repository (see [Models](#models)) |
>
> This repository is no longer under active development. It remains available for anyone who needs the original v0.9.x models, for example to reproduce published results or maintain existing workflows. Thank you to everyone who built with LTX-Video.

<div align="center">

# LTX-Video (Legacy)

**The original LTX video generation model (v0.9.x)**

[![Website](https://img.shields.io/badge/Website-LTX-181717?logo=google-chrome)](https://ltx.io/model/ltx-2-5)
[![Model](https://img.shields.io/badge/HuggingFace-Model-orange?logo=huggingface)](https://huggingface.co/Lightricks/LTX-Video)
[![Demo](https://img.shields.io/badge/Demo-Try%20Now-brightgreen?logo=vercel)](https://app.ltx.studio/ltx-2-playground/t2v)
[![Paper](https://img.shields.io/badge/Paper-arXiv-B31B1B?logo=arxiv)](https://arxiv.org/abs/2501.00103)
[![Trainer](https://img.shields.io/badge/LTXV-Trainer-9146FF?logo=github)](https://github.com/Lightricks/LTX-Video-Trainer)
[![Discord](https://img.shields.io/badge/Join-Discord-5865F2?logo=discord)](https://discord.gg/ltxplatform)

</div>

---

## Table of Contents

- [Introduction](#introduction)
- [What's New](#news)
- [Models](#models)
- [Quick Start Guide](#quick-start-guide)
  - [Online demo](#online-inference)
  - [Run locally](#run-locally)
    - [Installation](#installation)
    - [Inference](#inference)
  - [ComfyUI Integration](#comfyui-integration)
  - [Diffusers Integration](#diffusers-integration)
- [Model User Guide](#model-user-guide)
- [Community Contribution](#community-contribution)
- [Training](#training)
- [Control Models](#control-models)
- [Join Us!](#join-us)
- [Acknowledgement](#acknowledgement)

# Introduction

LTX-Video is the first DiT-based video generation model that contains all core capabilities of modern video generation in one model: synchronized audio and video, high fidelity, multiple performance modes, production-ready outputs, API access, and open access. It can generate up to 50 FPS videos at native 4K resolution with synchronized audio in one pass.
The model is trained on a large-scale dataset of diverse videos and can generate high-resolution videos with realistic and diverse content.

The model supports image-to-video, multi-keyframe conditioning, keyframe-based animation, video extension (both forward and backward), video-to-video transformations, and any combination of these features.

### Image-to-vid