Azure Cosmos DB and Azure SQL Database are both widely used database services offered by Microsoft Azure. While they serve the purpose of storing and managing data, there are several key differences between these two services that determine when and how they should be used.

1. Scalability and Performance: Azure Cosmos DB is designed to scale horizontally across multiple regions and provide global distribution of data with low latency. It offers limitless elastic scalability and can handle massive amounts of throughput. On the other hand, Azure SQL Database is built for vertical scalability within a single region and has certain limitations on scalability and performance.

2. Data Model: Azure Cosmos DB follows a NoSQL document data model, where data is stored in flexible JSON-like documents. It supports schema-less data and can handle unstructured, semi-structured, and structured data. In contrast, Azure SQL Database follows a relational data model, where data is organized into tables with predefined schemas that enforce data integrity and relationships.

3. Consistency Models: Azure Cosmos DB provides several consistency models, including strong, bounded staleness, session, and eventual consistency. This allows developers to choose the level of consistency that best suits their application requirements. Azure SQL Database, on the other hand, offers strong consistency by default, ensuring that every read operation retrieves the latest committed data.

4. Querying Capabilities: Azure Cosmos DB comes with a powerful querying capability using SQL-like syntax, which allows developers to perform complex queries on the stored documents using a familiar language. It also provides support for NoSQL-specific operations like filtering, sorting, and projecting on nested properties. In contrast, Azure SQL Database supports traditional SQL queries that are optimized for relational data and come with powerful querying capabilities for joins, aggregations, and complex filtering.

5. Multimodel Support: Azure Cosmos DB is a multimodel database that supports multiple APIs including SQL, MongoDB, Cassandra, Gremlin, and Table. This means that developers can choose the API that suits their application and work with their preferred programming model. Azure SQL Database, on the other hand, primarily supports the relational SQL API and is focused on supporting relational data storage and retrieval.

6. Cost Model: Azure Cosmos DB is billed based on provisioned throughput and consumed storage, which allows for precise control over the resources used and can be cost-effective for applications with varying workload patterns. Azure SQL Database, on the other hand, is billed based on the compute and storage resources provisioned, which may result in higher costs for fluctuating workloads or for applications that require high scalability.

In Summary, Azure Cosmos DB and Azure SQL Database differ in scalability, data models, consistency models, querying capabilities, multimodel support, and cost models, enabling developers to select the most suitable database service for their specific needs.