[
LTX-2 Model
]
Introducing LTX-2 model
A new standard for AI video generation
LTX-2 Model
A new standard for AI video generation
LTX-2.5 Model
A stronger foundation for the worlds already being built on LTX. Native multishot, precise editing, and 4K HDR output, built to hold up from first draft to final render. Learn More →
Generate connected scenes, not single clips. Hold character, environment, lighting, and voice consistent across wide, medium, and close-up shots in one generation.
Let the model set the pace. Clip length is predicted from the described action, so scenes land at the right duration without manual tuning.
Generate high-resolution HDR footage built for professional finishing. Output drops straight into your grading and color pipeline, ready for the big screen.
Generate every scene from a grid of high-fidelity keyframes that focus detail where it matters most. Get industry-leading pixel quality that holds up frame by frame, even on a cinema screen.
Creative control that holds up under pressure. Structure, motion, camera behavior, and identity can be directed with intent rather than guessed by the model.
Depth-aware generation
OpenPose driven motion
Camera control
Stylistic and visual consistency
Audio to video
The model adapts to your worlds, characters, and creative DNA. Customization becomes part of the workflow, not a research project.
LoRA training support
Style LoRAs
Tools for upscaling, restoration, and detail recovery, powered by the model’s multi-scale rendering pipeline.
Detail upscaling
Recreate and generate elements of already existing videos. Edit with surgical precision.
Retake
Extend scene
Low-cost, high-speed generation for rapid iteration, storyboards, and previews.
Technical characteristics:
Higher detail and motion stability for final, commercial-grade renders.
Technical characteristics:
Research
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.
-3_2%20-%20Dark.webp)
Asymmetric dual-stream DiT architecture. Joint audio-video generation with bidirectional cross-attention and modality-aware classifier-free guidance. Open weights and code.
LTX-2.5 is built on a 22B-parameter asymmetric dual-stream diffusion transformer. The weights and code are publicly available.
For academic teams pushing the boundaries of video generation, world simulation, and multimodal AI. Grants, model access, and research partnerships with the LTX team.
FAQs
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.
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.
Yes. LTX-2 is available for download in supported configurations for local and private deployment.
Yes. The open-source version is available on GitHub and Hugging Face, enabling customization, fine-tuning, and local execution.
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
Supported variants include LTX-2-fast and LTX-2-pro, optimized for speed or maximum quality.
Yes. Developers can experiment with the model via the LTX-2 Playground.
Yes. LTX-2 supports integration with tools such as ComfyUI and Fal for custom pipelines and experimentation.
