Technology Blog Posts by SAP
cancel
Showing results for 
Search instead for 
Did you mean: 
Yogananda
Product and Topic Expert
Product and Topic Expert
786

 

Understanding Serving Endpoints in SAP Databricks

The below screenshot highlights the "Serving endpoints" section of SAP Databricks, a feature designed to deploy and manage machine learning models seamlessly. This interface allows users to experiment with various large language models (LLMs) and connect them securely to external applications. Key models like Claude Sonnet 4.5, GPT-OS 120B, and OpenAI GPT-5 are showcased, indicating the platform's support for cutting-edge AI technologies.

2025-10-19_22-56-18.png

What Are Serving Endpoints?

Serving Endpoints in Databricks allow users to expose machine learning models as REST APIs, making it easy to integrate them into applications, dashboards, or other services. Whether you're working with open-source models or proprietary ones, Databricks provides a seamless interface to manage them.

 

  • Featured Models: At the top, we see high-profile models like:

    • Claude Sonnet 4.5
    • GPT OSS 120B
    • OpenAI GPT-5
    • GPT OSS 20B

    These models are either hosted directly on Databricks or integrated externally, and users can quickly Use, Copy, or Configure them.

2025-10-19_23-06-54.png

  • Model Table Overview: Below the featured section, a detailed table lists various models with metadata such as:

    • State: All models are marked "Ready" for deployment.
    • Served Entities: Indicates the backend system or model variant.
    • Tags & Tasks: Most models are tagged for Chat or Embeddings, with tasks primarily focused on conversational AI.
    • Created By & Last Modified: Useful for tracking ownership and freshness of models.

Playground examples

2025-10-19_23-10-39 (1).gif