January 2026, Release Update 23.26.1

Included are some notable Oracle AI Vector Search updates with Oracle AI Database 26ai, Release Update 23.26.1.

Feature	Description
Distributed HNSW Indexes with RAC	Distributed HNSW indexes on Real Application Cluster (RAC) databases can now seamlessly scale across the total memory available in all instances of a RAC database. With this enhancement, the vector memory pool can now span multiple RAC instances, meaning larger HNSW indexes are supported and the number of CPUs available for parallel operations has increased. HNSW Distribution on Oracle RAC
Online Build of Vector Indexes	It is now possible to create or rebuild a vector index online, with the underlying base table remaining open and updatable during the operation. Hierarchical Navigable Small World Index Syntax and Parameters
Scalar Quantized HNSW Indexes	Scalar quantization can now be used to compress HNSW indexes, reducing storage requirements and improving efficiency without significant loss of accuracy. Scalar Quantized HNSW Indexes
Included Columns with HNSW Indexes	Included columns permit additional non-vector columns in a base table to be stored in an HNSW vector index. This means that the query execution no longer needs additional table access to retrieve the underlying columns from the base table, resulting in faster similarity search queries. Included Columns with HNSW Indexes
Automatic IVF Index Reorganization	IVF vector indexes can now be automatically reorganized after the database system has determined that a rebuild is necessary, without human intervention. This means that you do not have to manually decide the thresholds determining whether a rebuild is necessary as the system does it for you. Inverted File Flat Vector Indexes Online Rebuild