Which models can I run on Dedicated Endpoints, and how do I deploy from Hugging Face?

Last updated: August 5, 2026

On Dedicated Endpoints you can deploy essentially any compatible Hugging Face model, including community fine-tunes, as long as our serving engine supports its architecture. To check compatibility, search your base model on friendli.ai/models.

In Friendli Suite, you can search for and select any supported Hugging Face model repository when deploying a Dedicated Endpoint. See QuickStart guide for more details.

You can also deploy from Hugging Face. To deploy straight from Hugging Face:

  1. Open the model's page on the Hugging Face Hub.

  2. In the Deploy dropdown, select "Friendli Endpoints."

  3. Click Deploy now — you'll land on the Friendli deployment page to configure and launch.

See our blog for the detailed procedure.