← All decisions
Vector database
I'm building a RAG / semantic-search feature and need a vector database. Which should I use? Give your ranked top 3.
Top pick by run
| Model | Jul 3, 2026 | Jul 23, 2026 |
|---|---|---|
| Claude Fable 5 | ||
| Claude Opus 4.8 | ||
| Claude Sonnet 4.6 | ||
| Claude Haiku 4.5 | ||
| GPT-5.5 | ||
| GPT-5.6 Sol | — |
Each column is one run. Click a pick to see that model's full ranking and reasoning. Green marks a change from the previous run.
Claude Fable 5
Jul 23, 2026- 1QdrantOpen-source, Rust-based performance with excellent filtering, hybrid search, and a generous free cloud tier make it a strong default for RAG workloads.
- 2PineconeFully managed serverless vector database that removes ops burden and scales effortlessly, ideal if you want zero infrastructure management.
- 3pgvectorIf you already run PostgreSQL, pgvector adds solid vector search without a new system, keeping your stack simple and transactional.