Evidence-first authority page
Rerankers Explained: Why RAG Systems Use a Second Retrieval Pass
What a reranker does in a RAG pipeline, and why a two-stage retrieve-then-rerank approach often outperforms retrieval alone.
Direct answer
A reranker is a second-stage model that re-scores an initial set of retrieved documents for relevance to the specific query, applied after a faster initial retrieval step narrows down candidates from a much larger corpus — this two-stage approach balances speed (fast initial retrieval across everything) with accuracy (a more expensive but more accurate reranking model applied only to the smaller candidate set).
Last reviewed 2026-08-29
This page remains noindex while its page-level source receipts are completed and verified. It is available for review, not presented as publication-ready authority.
Why a two-stage approach rather than one
Running a highly accurate but slow ranking model across an entire large corpus for every query would be too slow for real-time use — retrieving a smaller candidate set quickly first, then reranking just that smaller set with a more accurate model, gets most of the accuracy benefit at a fraction of the computational cost.