LTX-2.3 – AI Video Generator
LTX-2.3 is Lightricks’ open-weight AI video model, designed for local production, rapid iteration, longer clips, native audiovisual generation, vertical video, and advanced creator workflows through tools such as ComfyUI
LTX-2.3 is Lightricks’ open-weight AI video model, designed for local production, rapid iteration, longer clips, native audiovisual generation, vertical video, and advanced creator workflows through tools such as ComfyUI

Best for: Open-weight local video production
Quality: Good
Speed: Varies by platform
Pricing: Low
Audio: Yes
Dialogue: Yes
Lip sync: Yes
Resolution: 1080p, 1440p, 4K
Max duration: Up to 20 seconds at 1080p; up to 10 seconds at 1440p/4K
Local use: Yes
Developer: Lightricks
AI video modes: Text-to-video / Image-to-video / Video-to-video / Video extension / Video editing / Motion transfer
Open-weight and local workflows, technical creators, ComfyUI users, rapid iteration, longer clips, native audiovisual generation, vertical video, and projects requiring greater control over the production pipeline. Its biggest advantages are low cost, speed, local flexibility, long-duration support, and an unusually complete creator toolset. It is currently one of the strongest open-weight options, but its raw visual quality, complex physics, emotional performance, and temporal stability generally remain behind the leading proprietary models.
Good — Follows clear camera, lighting, atmosphere, and motion instructions reliably, particularly when the prompt describes one focused shot. Complex scenes with many simultaneous demands remain less dependable.
Good (provisional) — Produces stable results in focused shots with one main subject and controlled movement. Complex scenes, fine textures, multiple characters and longer generations remain more likely to develop visual drift or temporal degradation.
Advanced — Supports prompt-based direction plus official camera-motion LoRAs for dolly in, dolly out, dolly left, dolly right, jib up, jib down and static shots. IC-LoRA workflows can transfer camera movement and scene structure from a reference video. LTX Studio exposes additional camera presets, although advanced control is strongest through Studio or ComfyUI workflows.
This model can be used directly through LTX’s official sites/platforms, downloaded for local use, or accessed through selected all-in-one AI platforms.
Easiest: LTX Studio
For creators who want a visual, no-code production suite to generate, edit, and assemble videos using LTX-2.3.
Advanced: LTX Developer Console
For developers and advanced users who want direct model access, testing tools, and more control over production workflows.
Local: LTX-2.3 open weights
For users who want to download and run the model on capable local hardware using tools such as ComfyUI. Local setup requires substantial GPU resources and technical familiarity.
We usually prefer all-in-one AI platforms for accessing models like LTX-2.3. One subscription gives you access to many of the most popular AI image and video models, making it easy to compare them side-by-side and quickly switch between them. This can save both time and money, even when the cost per generation is sometimes slightly higher.
The following are the all-in-one AI platforms we currently use and recommend: