# Agentic RAG Router LLM

**URL:** <https://community.deeplearning.ai/t/agentic-rag-router-llm/872372>\
**Category:** Retrieval Augmented Generation\
**Tags:** week-module-4, coursera-platform\
**Created:** [August 13, 2025, 11:26am UTC](https://community.deeplearning.ai/t/agentic-rag-router-llm/872372 "2025-08-13T11:26:15Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![MDF1](https://avatars.discourse-cdn.com/v4/letter/m/c67d28/32.png) [@MDF1](https://community.deeplearning.ai/u/MDF1)\
**Post date:** [August 13, 2025, 11:26am UTC](https://community.deeplearning.ai/t/agentic-rag-router-llm/872372/1 "2025-08-13T11:26:15Z")

</div>

About the Router LLM: How can it decide when information retrieval from the RAG system is necessary, when the final response is generatet from a separate LLM. Does the Router LLM have the same knowledge as the generator LLM?

---

<div class="post-metadata">

**Author:** ![Girijesh](https://sea1.discourse-cdn.com/dlai/user_avatar/community.deeplearning.ai/girijesh/32/37283_2.png) [@Girijesh](https://community.deeplearning.ai/u/Girijesh)\
**Post date:** [August 13, 2025, 3:18pm UTC](https://community.deeplearning.ai/t/agentic-rag-router-llm/872372/2 "2025-08-13T15:18:06Z")

</div>

Dear @MDF1,

Welcome back to the Community!

In a RAG system, the **Router LLM** decides if a query needs retrieval before the Generator LLM answers.

- It **doesn’t have to have the same knowledge** as the Generator — often it’s smaller/faster.

- Decision is based on **prompted rules, confidence checks, and domain keywords**.

–  
_Keep Learning AI with DeepLearning.AI_ - Girijesh