Amazon S3 Vectors increases similarity search results to 10,000
Amazon S3 Vectors now supports up to 10,000 similarity search results per query, a 100x increase from the previous limit. This enhancement is particularly useful for complex retrieval pipelines that require further processing to refine results. Engineers using the latest AWS SDK can update their applications to specify the higher limit and receive paged results.
- →Increase similarity search results limit
- →New result pagination and pricing details
Enhancements (1) ›
- Increase similarity search results limit
Amazon S3 Vectors now supports returning up to 10,000 similarity search results per query, a significant increase that aids in retrieving a more comprehensive set of candidates. This change is valuable for applications utilizing multi-stage retrieval pipelines that involve reranking, aggregations, or deduplication. Users will need to update their applications with the latest AWS SDK and specify the new limit in their QueryVectors API requests.
Notes (1) ›
- New result pagination and pricing details
Query results are now returned across multiple pages, allowing for immediate processing of the first page while subsequent pages are retrieved. A small data-returned fee applies for results exceeding 512 KB per query, with the initial 512 KB being free. This update is available in all AWS Regions where S3 Vectors is offered.
https://aws.amazon.com/about-aws/whats-new/2026/06/s3-vectors-supports-10000-search-results-per-query
Related releases
- Amazon S3 enhances access denied errors with specific policy ARNs AWS What's New ·
- AWS Clean Rooms now supports exporting privacy-enhanced analysis logs for SQL queries AWS What's New ·
- AWS Security Blog Details How to Secure S3 Buckets AWS Security Blog ·
- AWS Glue integrates with SageMaker Unified Studio for one-click data access AWS What's New ·
- AWS Backup for S3 adds direct read-only access to backup data AWS What's New ·
- Amazon MSK adds authorizer log delivery for provisioned clusters AWS What's New ·