How Agent Hub Uses Large Language Models (LLMs)
Agent Hub uses a configured large language model (LLM) to support orchestration and provide direct responses to user requests.
When you enter a prompt without selecting an agent, the orchestrator uses the configured LLM to help determine how to handle your request. If your request doesn't require a specialized agent or tool, Agent Hub can use the LLM to respond directly.
An LLM and an agent aren't the same thing. An LLM provides general language-understanding and generation capabilities, while an agent is designed to perform specific tasks using specialized instructions, tools, and other capabilities. Agents can also use their own configured LLMs.
The models available in Agent Hub depend on the selected processing region. An Agent Hub administrator selects the processing region and an available model for orchestration and direct responses.
For information about how Agent Hub processes requests, see How Agent Hub Processes Requests.
For information about configuring the processing region and model, see Understand Agent Hub Model Configuration.