🎯 Prefix caching

The items that name it 3

• Introducing AnySpeechRecognizer and AnyDecisionModel: Open Models for Speech and Decisions on Apple Platforms
  article2026-10-29
• Cutting our inference bill by 68% by being obsessive about optimization
  writeup2026-10-19

  Reuses already computed prompt prefixes across requests. Cloudflare split the system prompt at a marker so the part shared by all repositories got its own cache breakpoint, and the first request of a session could read from the cache.

• The Roadmap to Mastering LLM Inference Optimization
  article2026-09-21