LTX 2.5 Video Model Released: Native Multi-Shot Generation, ComfyUI Support

linoy_tsaban · x · 2026-08-12

Lightricks has released the new open-source video model LTX 2.5, featuring 22B distilled and full checkpoints. The update introduces a new diffusion decoder that maintains fast inference speeds while reducing compression artifacts. It also supports native multi-shot generation, allowing users to create multiple connected shots in a single pass. Day-0 integration with 🧨Diffusers and ComfyUI is included.

Related event: LTX-2.5 Open-Source Video Model Released: Native Multishot and Fast Inference(20 posts)→

Original post →

More from Multimodal

Multimodal channel →