OpenAI adds 12-hour prompt caching for API agents
The Responses API option sets a 12-hour minimum cache lifetime for shared prompts, at twice the cache-write rate of the 30-minute option.
By Ryan Merket · Published
Primary source: X
Why it matters
The 12-hour option targets a recurring cost in persistent agents: resending the same long prompt after an idle gap. Its 2.5x cache-write rate means the savings depend on measured reuse, not just keeping an agent active.
OpenAI added an optional 12-hour prompt-cache lifetime to its Responses API on October 20th, giving developers a way to reuse stable instructions and tool definitions across requests separated by hours. The change is aimed at persistent agents that repeatedly send the same large context, including OpenAI's own dots product, according to the OpenAI Developers account.…
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.