Which model runs in LTX Desktop?

LTX Desktop generates video using the LTX-2.3 model — the same open-weight model available on Hugging Face and used across the LTX platform.

LTX-2.3

LTX-2.3 is Lightricks' open-weight video generation model. In LTX Desktop it's available in two variants:

  • LTX-2.3 Fast — optimised for speed; good for iteration and drafts
  • LTX-2.3 Pro — higher quality output; slower to generate

In local mode, the model weights are downloaded to your machine on first launch — you can download LTX Desktop to get started. In API mode, generation runs on Lightricks' infrastructure using the same model.

Open weight

LTX-2.3 is an open-weight model, meaning the weights are publicly available. This is what makes local generation possible — you're running the actual model on your own hardware, not a cloud service. You can find the model weights on Hugging Face.

LoRAs and custom workflows

Because LTX Desktop is built on an open-weight model, it supports community LoRAs and custom model workflows. See LoRAs and custom workflows in LTX Desktop for more.

 

 

We'd love your feedback. Did this article help?