GoKawiilNew open-source index PRISM claims to beat pgvector's IVFFlat and HNSW
Developers have released PRISM, an alpha-stage PostgreSQL index access method built in C that targets approximate nearest neighbor search. It combines hierarchical centroid routing, RaBitQ quantization and SIMD-optimized distance calculations, and supports dynamic inserts, updates and deletes alongside existing pgvector columns.
GoKawiil's interpretation of the reporting above, not reported fact.
The developers claim PRISM outperforms pgvector's IVFFlat and HNSW on both query speed and index build time, including on a 100-million-vector dataset using a quarter of the memory HNSW requires. If those claims hold up in independent testing, it could lower the hardware cost of running large-scale vector search directly inside Postgres rather than a separate vector database, though as an alpha release it is explicitly not yet meant for production use.
- PRISM is a new PostgreSQL extension for approximate nearest neighbor search, currently in alpha
- It claims faster queries and index builds than pgvector's IVFFlat and HNSW, using less memory at scale
- It supports dynamic updates and can index existing pgvector 'vector' and 'halfvec' columns directly
Source: github.com, 2026-10-14
Published there as: “Pg_vectorsearch: An index that is faster than pgvector's IVFFlat and HNSW”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.