November 17, 2026 · AzureLaunched
Generally Available: Semantic Reranker in Azure Cosmos DB
Semantic Reranker in Azure Cosmos DB improves search relevance by reordering candidate results based on the meaning of the search text and surrounding context. It uses an AI model to evaluate candidate documents and promote the ones that best match the intent of a query or natural language question, helping you deliver more accurate, context-aware answers in search and retrieval-augmented generation (RAG) scenarios.
With Semantic Reranker in Azure Cosmos DB, you can add semantic reranking as a dedicated AI step in retrieval workflows for RAG and agent applications. This helps reduce the need to build and tune custom ranking pipelines while keeping Azure Cosmos DB as the operational data store for search and retrieval experiences.
Related updates
Generally Available: Azure Cosmos DB full-text search BM25 score projection
Public Preview: Multiple-language full-text search in Azure Cosmos DB for NoSQL
Qwen3 embedding and reranking models for retrieval are now available in Amazon SageMaker JumpStart · AWS
AI Search - Reranking and API-based system prompt configuration in AI Search · Cloudflare