Definition
BM25 is a ranking function that scores how well a document matches a search query based on term frequency and how rare each term is across the collection. It is a refinement of TF-IDF and remains a strong baseline for lexical, keyword-based retrieval. Most sparse retrieval systems still use it or a close variant.Why It Matters for AI Search
BM25 is the workhorse behind the keyword half of modern hybrid retrieval. It rewards documents that use a query’s exact terms, weighted so that rare, specific words count more than common ones. For content, precise terminology matters: the exact name, model number, or phrase a buyer uses is what a BM25 method matches, where a semantic method might paraphrase it away. A page that never states the specific term can lose the lexical path entirely.Related Terms
TF-IDF
See also
Sparse retrieval
See also
Keyword search
See also
Hybrid retrieval
See also
Semantic search
See also