LTXV 13B - Full Version Release ! Getting started with the new LTXV video gen ! 2025 UPDATE 2 !

The LTXV model was always the fastest, even at 1280x720 resolutions, it is still faster than WAN, Hunyuan and others when used on Local GPUs. The model was a research release containing only 2B parameters. The new model is much larger and contains 13B parameters. This should result in a clean base model for finetuning, time will tell.

https://github.com/MushroomFleet/DJZ-Workflows/tree/main/XODA-LTXV-Video-Pack/ltxv-13B-097

You Will Need:

Models

• LTXV 13B 0.9.7 Dev

https://huggingface.co/Kijai/LTXV/resolve/main/ltxv-13b-0.9.7-dev_fp8_e4m3fn.safetensors (NEW)

Place inside: ComfyUI\models\checkpoints
includes VAE

Clip

• T5XXL / T5XXL_FP8

https://huggingface.co/Comfy-Org/mochi_preview_repackaged/blob/main/split_files/text_encoders/t5xxl_fp16.safetensors

Place inside: ComfyUI\models\text_encoders

Upscalers

• Spatial Upscaler

https://huggingface.co/Lightricks/LTX-Video/resolve/main/ltxv-spatial-upscaler-0.9.7.safetensors

• Temporal Upscaler

https://huggingface.co/Lightricks/LTX-Video/resolve/main/ltxv-temporal-upscaler-0.9.7.safetensors
Place inside: ComfyUI\models\upscalers