product · May 17, 2026
Qdrant 1.18 Ships TurboQuant for Improved Vector Compression
Share the canonical public link.
Abdon Pijpelink announced Qdrant 1.18.0 on May 11, 2026, featuring TurboQuant that delivers twice the compression ratio of scalar quantization with similar recall and speed. The update adds memory monitoring via Web UI and API, named vectors management without collection recreation, and audit logging query endpoint. Benchmarks on arxiv-titles-instructorxl-embeddings dataset showed TQ4 recall at 0.9193 versus SQ at 0.9285 and F32 at 0.9419. Qdrant 1.18.0 is available on GitHub with full changelog and upgrade recommendations for Cloud and self-hosted users.
Spend governor blocked model creation: provider_circuit_open (lane=dev, provider=together)