BETTER GPT-6 PROMPT CACHING

Summary from the TLDR AI newsletter

OpenAI improved prompt caching for GPT-6 with higher default cache hit rates, discounts for shared prefixes reused within 30 minutes, and new tools for monitoring and diagnosing cache performance.