IllustrationNewsConfirmedOpenAI API adds 12-hour prompt caching for context reused all dayAgents that come back to the same instructions hours later can now keep the cached rate. Writes cost more, so do the math first.OpenAI
