Pyyan / Compare / Qdrant vs pgvector vs Milvus vs Weaviate
Qdrant vs pgvector vs Milvus vs Weaviate
4 of 5Vector Databases · verified 13 Aug 2026
| Specification | Qdrant | pgvector | Milvus | Weaviate |
|---|---|---|---|---|
| Summary | The fastest of the purpose-built stores. | Vectors inside the database you already run. | Built for billions of vectors. | Hybrid search with batteries included. |
| Index | HNSW | HNSW, IVFFlat | HNSW, IVF, DiskANN, GPU | HNSW |
| Hybrid search | Yes | Yes, with SQL | Yes | Yes, native |
| Hosting | Self-host and managed | Self-host or any managed Postgres | Self-host and managed | Self-host and managed |
| Licence | Apache 2.0 | PostgreSQL | Apache 2.0 | BSD-3 |
| p50 latency | 4ms | ~15ms | ~10ms | ~12ms |
| Category | Vector Databases | Vector Databases | Vector Databases | Vector Databases |
| Official | Qdrant ↗ | PostgreSQL ↗ | Zilliz ↗ | Weaviate ↗ |
Highlighted rows are where these differ.
Qdrant
- Written in Rust; 10 to 25% faster than Weaviate or Milvus on common workloads
- Strong filtering alongside vector search
- Self-host or managed, same engine
Best for latency-sensitive retrieval.
pgvector
- One database for vectors, rows and joins
- The honest default: reach for something else only when this stops working
Best for almost everyone, until scale says otherwise.
Milvus
- Distributed architecture with GPU index support
- Heavier to operate than the alternatives
Best for very large corpora.
Weaviate
- Vector, BM25 and metadata filters in one query
- Modules for embedding generation built in
Best for hybrid keyword and semantic search.