Which consumption pattern requires call states when working with an API?

Paging API calls break large result sets into smaller chunks but do not inherently introduce complex call state tracking beyond managing the current page.

Asynchronous API calls do not block the calling process, meaning the response is not immediately available; instead, the system continues other tasks and requires mechanisms like callbacks or polling to track the eventual completion and state of the original request. This necessity for tracking the request's progress or outcome introduces the need for managing call states.

Batch API calls group multiple requests into one, which streamlines communication but doesn't necessarily introduce the need for complex call state management for the batch itself once submitted.

Synchronous API calls block the calling process until a response is received, eliminating the need to track an ongoing 'call state' as the process waits for completion.