OpenAI adds 12-hour prompt caching option to its API

AIOpenAI says the API now offers 12-hour prompt caching, letting developers keep shared instructions, tools, and context cached for hours between requests. Cached input is billed at the lower cached-input rate even when requests are hours apart, which suits persistent agents that reuse context throughout the day.

AIOpenAI says developers can now choose 12-hour prompt caching in the API, which keeps shared instructions, tools, and context cached longer, even when requests are hours apart. The longer cache keeps cached-input pricing available across that window, and 12-hour cache writes cost 2.5x the input token price.