{"slug": "meet-leann-an-open-source-vector-index-that-shrinks-a-60m-chunk-rag-index-from-6", "title": "Meet LEANN: An Open-Source Vector Index that Shrinks a 60M-Chunk RAG Index from 201 GB to 6 GB", "summary": "UC Berkeley researchers released LEANN, an open-source vector index that stores a pruned graph and recomputes embeddings at query time, shrinking a 60-million-chunk RAG index from 201 GB to 6 GB. LEANN matches HNSW accuracy at 90% recall using 4 GB instead of 188 GB, at the cost of seconds-level search latency.", "body_md": "UC Berkeley’s LEANN vector index stores a pruned graph and recomputes embeddings at query time. It matches HNSW accuracy at 90% recall using 4 GB instead of 188 GB, at the cost of seconds-level search.\n\nThe post [Meet LEANN: An Open-Source Vector Index that Shrinks a 60M-Chunk RAG Index from 201 GB to 6 GB](https://www.marktechpost.com/2026/10/27/leann-vector-index-mlsys-2026/) appeared first on [MarkTechPost](https://www.marktechpost.com).", "url": "https://wpnews.pro/news/meet-leann-an-open-source-vector-index-that-shrinks-a-60m-chunk-rag-index-from-6", "canonical_source": "https://www.marktechpost.com/2026/10/27/leann-vector-index-mlsys-2026/", "published_at": "2026-10-27 21:54:37+00:00", "updated_at": "2026-10-27 23:41:32.286928+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-research", "ai-tools", "ai-infrastructure", "mlops"], "entities": ["UC Berkeley", "LEANN", "HNSW", "MarkTechPost"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/meet-leann-an-open-source-vector-index-that-shrinks-a-60m-chunk-rag-index-from-6", "markdown": "https://wpnews.pro/news/meet-leann-an-open-source-vector-index-that-shrinks-a-60m-chunk-rag-index-from-6.md", "text": "https://wpnews.pro/news/meet-leann-an-open-source-vector-index-that-shrinks-a-60m-chunk-rag-index-from-6.txt", "jsonld": "https://wpnews.pro/news/meet-leann-an-open-source-vector-index-that-shrinks-a-60m-chunk-rag-index-from-6.jsonld"}}