OpenAI adds 12-hour prompt caching to the Responses API, aimed at persistent agents like dots

#What happened

OpenAI added an optional 12-hour prompt-cache lifetime to its Responses API on October 20, 2026, announced by the OpenAI Developers account on X: "You can now choose 12-hour prompt caching in the API. Persistent agents, like dots, reuse context throughout the day." The option sets a 12-hour minimum cache lifetime for shared prompts, at twice the cache-write rate of the 30-minute option, so developers can reuse stable instructions and tool definitions across requests separated by hours.

#Why it matters

Persistent agents repeatedly send the same large context; the 12-hour option targets the cost of resending long prompts after an idle gap. OpenAI explicitly named its own dots product as a beneficiary of the feature.

Verified Oct 20, 2026.