[

LTX-2 Model

]

Introducing LTX-2 model

A new standard for AI video generation

Try LTX-2 Now

LTX-2.3 is here. All of LTX-2's
power, taken further

Beta

HDR Output

Delivered as an IC-LoRA on LTX-2.3. Generate directly in HDR or convert existing SDR footage to EXR. More grading latitude, more range, ready for real finishing pipelines.

Try LTX models

Retake

Regenerate just one section — replace video, audio, or both — without re-rendering the whole clip.

Try LTX models

Reframe

Repurpose a video to any aspect ratio for social, filling new areas with generated content.

Try LTX models

Audio to Video

Generate video where voice, music, and sound effects define structure, pacing, and motion.Built for production-grade workflows that require precise, harmonious control over audio-led scenes - from podcasts and avatars to voice-driven clips -not one-off demos or talking heads.

Try LTX models
Capabilities
Configurations
Research

Precise control over motion, structure, and camera

Creative control that holds up under pressure. Structure, motion, camera behavior, and identity can be directed with intent rather than guessed by the model.

Try LTX-2 Now

Depth-aware generation

Prompt: A cinematic wide aerial shot of a rugged desert mountain range at golden hour. A towering sandstone peak catches warm orange light, overlooking a vast arid basin and layered rocky hills under a soft, hazy pastel sky.

OpenPose driven motion

Prompt: A 3D animated medium shot of a youthful, blond astronaut in a red spacesuit standing inside a sleek white shuttle. He gazes confidently toward a large panoramic window revealing a vibrant galaxy, rendered with soft, cinematic lighting and a warm, polished aesthetic.

Camera control

Static
Dolly in
Dolly out
Dolly left

Stylistic and visual consistency

Audio to video

Customize the model to your data and identities

The model adapts to your worlds, characters, and creative DNA. Customization becomes part of the workflow, not a research project.

Try LTX-2 Now

LoRA training support

Style LoRAs

Prompt: A stop-motion style scene featuring birds made of yellow felt. One bird approaches a birdhouse and shares a worm with another, showcasing the tactile textures of wool, cardboard, and twine in a miniature set.
Prompt: An extreme close-up in a handcrafted aesthetic, focusing on a bird's head. The shot emphasizes the fuzzy, fibrous texture of the felt material and a simple black bead eye, mimicking a macro lens on a physical miniature.
Prompt: A miniature-world animation where a yellow felt bird stands in its nest while others pop out of colorful houses. The style uses "low-fi" textures like rough fabric and painted wood to create a charming, handmade look.
Prompt: A small yellow felt bird "flies" through a cardboard and fabric village. The animation mimics a side-scrolling effect using physical props, maintaining a consistent tactile, stop-motion appearance.

Restore and enhance video quality

Tools for upscaling, restoration, and detail recovery, powered by the model’s multi-scale rendering pipeline.

Try LTX-2 Now

Detail upscaling

Edit video without starting over

Recreate and generate elements of already existing videos. Edit with surgical precision.

Try LTX-2 Now

Retake

Original
Prompt: A live-action cinematic shot of a living room completely engulfed in roaring flames. The camera pans across the scene as fire consumes the bookshelves and furniture while embers fly through the smoke-filled air.
Prompt: A live-action horror style, thick black oily tentacles slither across the floor and furniture of a living room. Dark, glowing eyes peer through a heavy mist, creating an atmosphere of eerie supernatural dread.
Prompt: A live-action scene of a calm adult dog sitting in a living room, surrounded by a dozen energetic puppies. The puppies bark and run in circles around the patient dog in a lively and chaotic family moment.

Extend scene

Two tiers, one API.
Start on Fast, ship on Pro.

Fast

Low-cost, high-speed generation for rapid iteration, storyboards, and previews.

Technical characteristics:

  • Resolutions: 1080p, 1440p, 4K
  • FPS: 24 / 25 / 48 / 50
  • Duration: up to 20s at 1080p, up to 10s at 1440p/4K
Start with Fast

Pro

Higher detail and motion stability for final, commercial-grade renders.

Technical characteristics:

  • Resolutions: 1080p, 1440p, 4K
  • FPS: 24 / 25 / 48 / 50
  • Duration: up to 10s · required for audio-to-video, retake, extend
Compare models
//

Research

Performance

Built on a distilled hybrid architecture, LTX-2 delivers significantly higher generation throughput without compromising visual fidelity. It outperforms smaller models like WAN 2.2 14B under identical settings, enabling faster iteration and high resolution video workflows on modern GPU hardware.

Read the Research Paper →

GitHub
HuggingFace
Join the Discussion
Data center performance (H100)
Step Per Minute:
LTX-2 vs. WAN 2.2 14B
~18x faster
LTX-2 demonstrates a clear performance advantage, delivering dramatically higher step throughput than WAN 2.2 14B under identical generation settings on H100, making high resolution, long sequence video generation fast and production ready.

Technical report

Asymmetric dual-stream DiT architecture. Joint audio-video generation with bidirectional cross-attention and modality-aware classifier-free guidance. Open weights and code.

Read the Paper

LTX-2.3

Rebuilt VAE for sharper detail. 4x larger text connector for tighter prompt adherence. Native portrait, cleaner audio, HDR output, and precise control over motion and camera.

Explore the Model

Academic collaboration

For academic teams pushing the boundaries of video generation, world simulation, and multimodal AI. Grants, model access, and research partnerships with the LTX team.

Apply
//

FAQs

What is LTX-2?

LTX-2 is a professional-grade video generation model built for long-form, high-fidelity output with precise creative control. It is designed to support real production workflows rather than short experimental clips.

What makes LTX-2 different from the LTXV model?

While the LTXV model focuses on creator-friendly generation and editing, LTX-2 is optimized for developers and production teams that require predictable performance, deeper control, and measurable specs.

Does LTX-2 offer a downloadable model for local use?

Yes. LTX-2 is available for download in supported configurations for local and private deployment.

Is LTX-2 available as an open-source model?

 Yes. The open-source version is available on GitHub and Hugging Face, enabling customization, fine-tuning, and local execution.

How much does LTX-2’s API cost?

Pricing for the LTX-2 API depends on usage, resolution, and frame rate. Refer to LTX-2 pricing for detailed cost information regarding AI video generation API usage

What video generation models are supported by LTX-2?

Supported variants include LTX-2-fast and LTX-2-pro, optimized for speed or maximum quality.

Does LTX-2 offer a free trial?

Yes. Developers can experiment with the model via the LTX-2 Playground.

Does LTX-2 integrate with ComfyUI or Fal?

Yes. LTX-2 supports integration with tools such as ComfyUI and Fal for custom pipelines and experimentation.

Home
LTX-2
Open foundation models for video, audio, and world simulation
Research
Interfaces
Open Source
Resources
Pricing
Company
© 2026 All rights reserved