On March 5th, 2025, the Lightricks team launched the highly anticipated LTX-Video 0.9.5, an open-source AI video generation model based on the Diffusion Transformer (DiT).
As the latest iteration of the world's first real-time, high-quality video generation model, LTX-Video 0.9.5 offers a new experience for tech enthusiasts, developers, and commercial users with its powerful feature updates and open commercial license. The Lightricks official account announced on the morning of the launch that the new version not only improves generation quality but also adds several practical features, marking a new milestone for open-source AI video creation.
According to the official tweet, one of the biggest highlights of LTX-Video 0.9.5 is its commercial license support. This means that businesses and individual developers can use the model in commercial projects, greatly expanding its application prospects. Simultaneously, the new version introduces keyframe conditional support, allowing users to generate content directly from the video's first frame, last frame, or any intermediate frame, providing greater flexibility for video extension and editing.
Furthermore, the model shows significant improvements in resolution and quality, supporting real-time generation at up to 768x512 resolution and 24 frames per second, while reducing artifacts for smoother, more natural visuals. The ability to generate longer video sequences further caters to complex narrative needs, while native integration with ComfyUI allows users to seamlessly integrate it into their existing workflows for easier operation.
LTX-Video 0.9.5 is now available for download on GitHub and Hugging Face, supporting environments such as Python 3.10.5 and CUDA 12.2. The community generally believes that this update solidifies LTX-Video's position in the AI video generation field, and future optimizations in generation speed and higher resolution are highly anticipated. As Lightricks stated, LTX-Video 0.9.5 pushes "open-source AI video creation to new heights," opening a door to the future for creators worldwide.