BGE Reranker Base

BGE Reranker Base is a cross-encoder reranking model developed by the Beijing Academy of Artificial Intelligence (BAAI), designed to improve search relevance by re-scoring and re-ordering candidate documents. Built on a BERT-style architecture, it takes a query-document pair as input and produces a relevance score, enabling more precise retrieval than bi-encoder approaches.

The model was trained on large-scale relevance datasets and supports both monolingual English and multilingual reranking tasks. It integrates easily into existing search pipelines, typically applied as a second-stage reranker after initial retrieval with embedding models like BGE.

BGE Reranker Base is part of the popular BGE (BAAI General Embedding) family, widely adopted in RAG systems and enterprise search applications.