Tutorials

ComfyUI Official LTX-2.3 Workflow Examples — Templates, Models, Paths

Canonical ComfyUI reference for LTX-2.3: built-in I2V/T2V templates, required checkpoints, LoRAs, latent upscaler, Gemma 3 text encoder, and the exact model storage layout.

By ltx workflow

Editor's Note: The official ComfyUI docs page for LTX-2.3. It's the canonical reference for the bundled workflow templates and where to put each file. Mirrored here so the model + path conventions are easy to look up alongside our setup guide.

Key features

  • Finer details: new latent space & updated VAE for sharper textures, cleaner edges, more precise visuals
  • 9:16 portrait support: much-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 (e.g. frozen frames) for smoother 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

Getting started

  1. Update ComfyUI to the latest version.
  2. Open Template Library > Video and pick any LTX-2.3 workflow.
  3. Follow the popup to download the required models and run the workflow.

Workflows

ComfyUI ships two built-in LTX-2.3 templates:

Image-to-video

Standard I2V pipeline. Attach your reference image to LTXVImgToVideoConditioning. Video and synchronized audio are generated in a single diffusion pass.

Text-to-video

T2V pipeline using the same checkpoint and text encoder. No image input — LTXVImgToVideoConditioning is omitted.

Both templates can also run on Comfy Cloud if you don't have a local GPU with enough VRAM.

Model downloads

Checkpoints (place in ComfyUI/models/checkpoints/)

NameDescription
ltx-2.3-22b-dev.safetensorsBF16 full model
ltx-2.3-22b-dev-fp8.safetensorsFP8 quantized model

LoRAs (place in ComfyUI/models/loras/)

Name
ltx-2.3-22b-distilled-lora-384.safetensors

Latent upscale models (place in ComfyUI/models/latent_upscale_models/)

Name
ltx-2.3-spatial-upscaler-x2-1.0.safetensors

Text encoders (place in ComfyUI/models/text_encoders/)

Name
gemma_3_12B_it_fp4_mixed.safetensors

Model storage location

ComfyUI/
├── models/
│   ├── latent_upscale_models/
│   │   └── ltx-2.3-spatial-upscaler-x2-1.0.safetensors
│   ├── checkpoints/
│   │   ├── ltx-2.3-22b-dev-fp8.safetensors
│   │   └── ltx-2.3-22b-dev.safetensors
│   ├── loras/
│   │   └── ltx-2.3-22b-distilled-lora-384.safetensors
│   └── text_encoders/
│       └── gemma_3_12B_it_fp4_mixed.safetensors

Resources

Sources

#ltx-2.3#comfyui#official#templates#install