LTX-Video 13B 0.9.8 (and distilled / 2B variants)
The first-generation LTX-Video family: a 13B DiT plus a distilled 13B and a 2B distilled model, all built for speed via multiscale rendering.
- Resolution
- 1216x704 native (multiscale rendering to higher resolutions)
- Weights
- Downloadable
- API
- No
- Commercial use
- Permitted subject to conditions
It is one of the few open video models fast enough to approach real-time on a single GPU. Handles text-to-video, image-to-video and keyframe conditioning.
Custom LTXV Open Weights License, not OSI. Free below $10M annual revenue; above it, commercial use outside a signed agreement is stated to be strictly prohibited.
Revenue threshold: free below $10,000,000 in annual revenue, chargeable above it.
- Free only below a revenue threshold. Free for entities with under $10M annual revenue including commercial use; above that a Commercial Use Agreement must be obtained from Lightricks.
- Depends on your plan tier. Earlier 0.9.x releases were distributed under OpenRAIL-M, so terms differ by version. Check the licence file shipped with your exact checkpoint.
Specification
- What it takes in
- text, image, video
- What it makes
- video
- Maximum duration
- not established
- Maximum resolution
- 1216x704 native (multiscale rendering to higher resolutions)
- Aspect ratios
- not established
- Controls
- text prompt, reference image, keyframe conditioning, video extension, multiscale rendering, FP8 quantized checkpoints, distilled few-step sampling, ComfyUI and diffusers integration
- Weights
- Downloadable
- API
- No
- Pricing
- Free open weights under Lightricks' LTXV Open Weights License. Self-hostable: the 2B distilled checkpoints run on modest consumer GPUs and the FP8 13B checkpoints reduce memory further; exact VRAM figures are not published per variant in the README. apiAvailable is set false because no first-party documented developer API was verified - programmatic access is via third-party hosts (fal, Replicate) and the LTX Studio product.
- Released or updated
- 2025-07
Maximum duration and supported aspect ratios are not established here. That means no page we could retrieve stated them, not that the model lacks them.
Strengths
- Fastest open video family; distilled variants are near real-time
- 2B distilled variant runs on light hardware
- Free commercial use under $10M revenue
- FP8 checkpoints officially published
- Trained on licensed stock footage
Weaknesses
- Per-version license files with differing terms (OpenRAIL-M vs LTXV Open Weights) are easy to get wrong
- $10M revenue threshold
- Superseded by LTX-2 for quality and audio
- Official docs do not state hard max resolution or duration
Best for
- Latency-sensitive open-weights video generation and interactive tools, or as a lightweight local model.
Where the sources disagree
Recorded rather than resolved. Picking the tidier number would hide the fact that the vendor's own pages do not agree.
- One summary of the GitHub README reported '4K native, 50 fps, 60 sec' for the 13B variants; the Hugging Face model card instead describes 1216x704 at 30 fps with a recommendation to stay under 257 frames. Treating the model card as authoritative and leaving hard maxima null. License naming also conflicts across sources (OpenRAIL-M vs LTXV Open Weights License) because it genuinely changed between versions.
What to understand before using it
Explainers naming the mechanisms this model's own documentation mentions.
MCP servers that front Lightricks
Servers in the media MCP directory whose own description or homepage names this vendor. Absence of a server here is not evidence that none exists: most publish no tool list at all.
-
Comfyui
MCP server + Claude Code plugin for ComfyUI: run workflows, generate images, manage models & VRAM.
-
Corent
Generate images and videos via the Corent API, auto model routing with provider fallback.
-
Fal
Create and edit high-quality media across various formats including images, videos, audio, and 3D models.
-
OpenArt
Generate AI images and video from any agent, with OAuth and no API keys.
-
Picsart GenAI
Generate and edit images, videos and audio with 150+ AI models.
-
Runcomfy
Remote MCP server for RunComfy Serverless API (ComfyUI): deployments and async inference.
See it from the vendor
Shotcraft hosts no sample output. Judging a model means watching its own showreel with your own eyes, and a clip re-hosted here would be a clip we chose.
Nearby in the directory
LTX-2.5
A 22B-parameter diffusion transformer that generates synchronized video and audio in a single model, using a Gemma-class 12B text encoder and separate
Video generation
Adobe Firefly Video Model
The Adobe Firefly Video Model generates video from a text prompt, optionally guided by an image used as the first or final frame. Adobe's Firefly Serv
Video generation
Allegro
A small text-to-video model comprising a 2.8B VideoDiT and a 175M VideoVAE, released under Apache-2.0 in October 2024. It generates 88-frame clips at
Video generation