Vector Database Comparison 2025 to 2026: Milvus vs Qdrant vs Weaviate vs pgvector

How the vector database landscape changed from 2025 to 2026, with real star growth.

## Vector Database Comparison 2025 to 2026 Vector databases matured fast between 2025 and 2026. Here is how the four leaders compare now, with star counts from the 2026-07-31 snapshot. ### The four leaders | Database | Stars (2026-07) | Focus | |----------|----------------|-------| | Milvus | 45,439 | Enterprise, billions of vectors | | Qdrant | 33,697 | Production Rust, rich filters | | Weaviate | 16,677 | Hybrid graph + vector | | pgvector | 22,417 | Postgres extension | ### What changed from 2025 1. **Hybrid search became table stakes** - keyword + vector in one query is now expected 2. **pgvector 0.8+ added HNSW improvements** - Postgres users no longer need a separate DB for most apps 3. **Disk-based indexes matured** - Qdrant and Milvus handle billions of vectors on NVMe without RAM blowups 4. **Managed cloud won** - Qdrant Cloud, Zilliz (Milvus), Weaviate Cloud dominate new deployments 5. **Agent-era workloads** - memory layers (mem0 62,202 stars) integrate with vector DBs ### Decision for 2026 - Small app, Postgres already there: pgvector - Serious product, want managed: Qdrant Cloud or Zilliz - Need hybrid search with typo tolerance: Typesense (26,385 stars) or Weaviate - Research/benchmarking: Milvus (its benchmark tool is industry standard) ### FAQ **Did anything get worse?** Self-managing a vector DB is harder in 2026 - indexes and scaling grew complex; managed services are the sane default. **Is Chroma still relevant?** Yes, for local dev and prototypes; not for production scale.