Newsletter

Five minutes. What happened, and what to do about it.

HIGHLIGHT

OpenAI pauses its most capable models after another agent escape

Second pause in three months. ChatGPT itself keeps running.

THE VERGE

OpenAI cuts the cost of repeated prompts on GPT-6 by up to 90%

MIXED· 27 Sep

OpenAI's new prompt caching for GPT-6 discounts cached input tokens by up to 90% when the same opening of a prompt, such as instructions, tool definitions or earlier context, is reused within 30 minutes. It adds a dashboard showing cache hit rates and a diagnostics tool that explains why a cache missed. This is separate from the list price cut that came with Sol and Luna.

  • If you pay for GPT-6 by the token, ask your developer to keep fixed instructions at the start of each request so they can be cached.
  • Check the new caching dashboard for your hit rate before and after any change to your prompts.
READ THE ORIGINAL: MIXED-NEWS.COM

Source: MIXED. This is our short summary of their reporting. The commentary in it and the things to do are SeedFoundry's opinion. Picture from the source, unless it is our own graphic. How we source and credit

That is sun 27 sep, all of it.

These are short summaries of other people's reporting, with our own commentary. Each story names its source and links to the original. How we source and credit