We built a real-world bench for vectorDBs

VectorDBBench 1.0 is Here!

We're excited to launch VectorDBBench 1.0, a major update focused on a better user experience and more realistic testing.

What's New in 1.0:

• 🚀 Redesigned UI: A brand new homepage and integrated analytics pages make it easier than ever to visualize and compare test results.
• 🏷️ Label-Filter Tests: Simulate real-world queries with new tests for filtered search (e.g., color == "red").
• 🌊 Streaming Scenarios: Measure search performance while data is actively being inserted to test "read-while-writing" stability.
• 🔬 New BioASP Dataset: Added a 1024-dimensional dataset (1M & 10M vectors) for high-dimensional performance testing.
• ⚙️ Flexible Custom Datasets: Enjoy more powerful options to configure datasets that mirror your own data.

We have also re-tested and updated all benchmark results for Milvus, ZillizCloud, ElasticCloud, QdrantCloud, Pinecone, and OpenSearch(AWS).

Check out the new features and updated results today!