LTX-Video in ComfyUI
Real-time compressed-latent video generator capable of rendering 24 fps clips faster than playback on mid-range RTX desktops. Requires 8 GB–12 GB VRAM and loads from models/checkpoints/.
Updated Oct 17, 2026 · 1 min read
- Model Family
- Lightricks
- VRAM Target
- 8 GB–12 GB
- ComfyUI Folder
- models/checkpoints/
- Precision
- BF16 / FP8 safetensors
- Default Sampler
- LTX scheduler, 20–25 steps
Where to place LTX-Video weights on Windows
Real-time compressed-latent video generator capable of rendering 24 fps clips faster than playback on mid-range RTX desktops.
Place the primary `.safetensors` or `.gguf` weight file inside `models/checkpoints/` inside your ComfyUI directory, or point extra_model_paths.yaml to your existing shared models drive.
After moving the file, press **R** on the ComfyUI Studio canvas (or click Refresh) so the loader dropdown lists the new checkpoint immediately without restarting the server.
Recommended loader nodes and sampler settings
- **Precision & Memory:** BF16 / FP8 safetensors. Cards at the lower end of 8 GB–12 GB should pair this model with FP8 or GGUF quantization and VAE Decode (Tiled).
- **Sampler & Scheduler:** LTX scheduler, 20–25 steps. Wire the model output into KSampler and connect the matching CLIP and VAE streams.
- **Installer:** If you have not set up the desktop environment yet, get the Windows build at /download.
LTX-Video Quick Reference for ComfyUI Windows
| Parameter | Recommended Setting | Notes |
|---|---|---|
| Target Directory | models/checkpoints/ | Refresh canvas with R key after copying |
| VRAM Tier | 8 GB–12 GB | Use --lowvram or GGUF Q4 on 8 GB cards |
| Weight Format | BF16 / FP8 safetensors | Never load untrusted .ckpt pickle files |
| Sampler Recipe | LTX scheduler, 20–25 steps | Wire through core KSampler node |
FAQs
Which folder does LTX-Video go into in ComfyUI?
Put LTX-Video inside `models/checkpoints/` in your ComfyUI installation, then press R on the canvas to refresh node dropdowns.
Can I run LTX-Video on an 8 GB NVIDIA GPU?
Yes — use FP8 or GGUF Q4/Q5 weights via ComfyUI-GGUF and keep tiled VAE decoding enabled so peak VRAM stays under 8 GB.
Run on your hardware. Free forever.
The full ComfyUI engine: open source, fast, extensible, and yours to run however you want. Download the desktop app and open this workflow on your own machine.
