October 2026, Release Update 23.26.4

Included are some notable Oracle AI Vector Search updates with Oracle AI Database 26ai, Release Update 23.26.4.

Feature	Description
Product Quantization Support for HNSW Indexes	HNSW indexes now support product quantization. Product quantization enables the compression of vectors into a more compact representation while preserving similarity characteristics. Product Quantized HNSW Indexes
Multi-Vector Search Using HNSW Indexes	HNSW indexes now support multi-vector similarity searches to improve search performance. Multi-Vector Search Using HNSW Indexes
GraphRAG	GraphRAG combines vector search with graph traversal to retrieve connected facts from across documents. GraphRAG
Snapshots Support for Distributed HNSW Indexes	Distributed HNSW indexes support snapshots to maintain performance and accuracy HNSW Distribution on Oracle RAC
Covering Columns Support for Distributed HNSW indexes	Distributed HNSW indexes support included columns. HNSW Distribution on Oracle RAC
Scalar Quantization Support for IVF Indexes	Global IVF indexes now support scalar quantization to reduce vector storage requirements Scalar Quantized IVF Indexes