Put the reusable prompt first
Explain why prompt caches reward stable prefixes and identify which prompt parts should move before dynamic request data.
The move: design the prefix, not just the prompt. Prompt caching rewards repeated beginnings. If the first large block changes because you put a ticket ID, date, retrieval chunk, or user profile at the top, you have made every call look new. If the first block is stable, the provider has a reusable shape to recognize. Stable prefix Keep the role, policy, rubric, tool definitions, examples, and long reference material together. These are the parts you expect to reuse across many requests. Treat this block like a versioned asset. Small edits can be valuable, but random churn makes hit rates…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in