OpenAI’s announcement on 22 September 2026 describes improvements to prompt caching in the forthcoming GPT‑6 model. The vendor claims the updates increase cache hit rates, provide new diagnostic tools, introduce explicit breakpoints, and add controls that together reduce inference latency and operational costs. No further technical details or performance metrics were provided.
← ALL NEWS
NEWS · 2026-09-24
OpenAI announces better prompt caching for GPT-6
OpenAI says its upcoming GPT-6 model includes enhancements to prompt caching that aim to raise cache hit rates, add diagnostics, explicit breakpoints, and controls intended to cut latency and costs.