LTX 2.5 is Here — Fastest Open-Source Video Model Yet in ComfyUI

In this video, I'll show you how to run LTX 2.5 — Lightricks' newest and fastest open-source audio-video model — inside ComfyUI. With native 2K output, a built-in Duration Predictor, and dramatically faster generation than LTX 2.3, this is the biggest jump the LTX family has seen. The workflow features a clean T2V/I2V toggle switch so you can switch between text-to-video and image-to-video without changing your setup.

Featuring LTX 2.5huggingface.co

What you'll learn

• Performance Upgrades — LTX 2.5 vs 2.3 speed and resolution comparison
• Hardware Efficiency — Generating up to 2K video on consumer GPUs
• Installation & Setup — Transformers, spatial upscalers & encoders
• ComfyUI Workflow — T2V and I2V with a single switch
• Duration Predictor & manual sigma adjustments for better motion
• Prompt Engineering for open-source video models
• Motion & Consistency — Strengths in lip-sync and fast-movement limits

Workflow

Workflow downloadbuymeacoffee.com

Model links

models/diffusion_models

BF16 Distilledhuggingface.co INT8 Distilled (ComfyUI)huggingface.co NVFP4 Distilledhuggingface.co BF16 Devhuggingface.co INT8 Dev (ComfyUI)huggingface.co

models/text_encoders

gemma4-12b-with-proj-ltx-2.5-bf16.safetensorshuggingface.co

models/vae

ltx-2.5-video-vae-bf16.safetensorshuggingface.co ltx-2.5-audio-vae-bf16.safetensorshuggingface.co

models/model_patches

Duration Headhuggingface.co

models/latent_upscale_models

ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensorshuggingface.co

• 👍 Support the Channel If you found this helpful:

• 👍 Like the video
• 🔔 Subscribe for more AI workflows
• 💬 Comment what you want to see next

#LTX25#ComfyUI#AIVideo#Lightricks#GenerativeAI#TextToVideo#ImageToVideo#OpenSourceAI