LTX-2.3 Day-0 support in ComfyUI: Enhanced Quality for Audio‑Video Generation
Enhanced quality for OSS audio-video generation
Hi community! We’re excited to announce that LTX-2.3, the latest evolution of Lightricks’ open-source audio-video generation model, is now natively supported in ComfyUI! Building on the foundation of LTX-2, this release delivers major quality improvements across fine details, portrait video, audio, image-to-video, prompt understanding, and text rendering.
Model Highlights
LTX-2.3 brings a comprehensive set of quality upgrades to the LTX family.
Finer Details: New latent space & updated VAE for sharper textures, cleaner edges, and more precise visuals.
9:16 Portrait Support: Greatly improved quality for vertical portrait videos, perfect for social media & mobile.
Better Audio: Cleaner sound with reduced noise, enhanced dialogue, music, and ambient audio.
Improved Image-to-Video: More consistent motion and fewer glitches, such as frozen frames, for smoother, more natural animations.
Smarter Prompt Understanding: Improved text encoder for more accurate interpretation of complex prompts.
Clearer Text Rendering: More accurate text and letter rendering in videos.
Example outputs
Image to Video
Text to Video
Getting Started
Update ComfyUI to the latest version (0.16.1) or visit Comfy Cloud
Access Workflows: Go to Template Library → Search → LTX-2.3
Download Models: Follow the prompts to download the required models
Start Creating: Configure your prompts and inputs, then run the workflow
As always, enjoy creating!
Please also mention the hardware/memory requirements so the community knows whether they can run the full model locally or if they’ll need to wait for quantized or distilled versions. Thanks for the early support and the blog post!