Amazon S3 Vectors
25/100S3-native vector storage & similarity search, pure pay-per-use.
aws.amazon.com// scorecard · 25/100
verified 1mo ago// rankings · cost per standard workload
Storage $0.18 + 1M queries $2.50 + data processed $6.14 = ~$8.82/mo (+$0.61 one-time PUT)
aws.amazon.com/s3/pricing// overview
Amazon S3 Vectors is the first cloud object store with native support to store and query vectors, delivering purpose-built, cost-optimized vector storage for AI agents, AI inference, and semantic search of content in Amazon S3. It reduces costs of uploading, storing, and querying vectors by up to 90% while supporting up to billions of vectors with sub-second query performance and the same elasticity and durability as Amazon S3. It integrates with Amazon Bedrock Knowledge Bases and OpenSearch Service for RAG and tiered vector search strategies.
Best for: Cost-optimized storage and querying of large-scale vector data for AI agents, semantic search, and infrequent-access workloads
// security & compliance
// features
- Dedicated APIs to store, access, and query vectors without infrastructure provisioning
- Up to 90% lower costs for uploading, storing, and querying vectors
- Store up to 2B vectors per index and 10,000 indexes per bucket (up to 20 trillion vectors)
- 100ms lowest warm query latency with strong consistency
- Native integration with Amazon Bedrock Knowledge Bases and OpenSearch Service
- Optimized for long-term vector storage with infrequent queries