Weaviate 1.31 Release: MUVERA, BM25, and HNSW Snapshotting

AI Impact Summary

Weaviate 1.31 introduces significant enhancements to its core functionality, primarily centered around multi-vector embeddings with MUVERA and refined keyword search operators. The ability to add new vectors to existing collections, coupled with HNSW snapshotting for dramatically reduced index startup times, represents a substantial performance boost, particularly for large deployments. This release expands model integrations and offers granular control over search queries, enabling greater flexibility and optimization for diverse use cases.