OpenAI adds 12-hour prompt cache to Responses API

On October 20 OpenAI added a "12h" TTL option to prompt caching in the Responses API alongside the default "30m", on GPT-5.6 and later models. Cache writes cost 1.25x the uncached input rate for 30 minutes and 2.5x for 12 hours, while reads cost 0.1x (0.05x on GPT-6.1 Sol). A single request can set different TTLs per context block.

• The "12h" TTL works on GPT-5.6 and later; the default stays "30m"
• Cache writes cost 1.25x input rate for 30 minutes and 2.5x for 12 hours
• Cache reads cost 0.1x, or 0.05x on GPT-6.1 Sol
• An unsupported TTL value returns a 400 invalid_request_error