Glossary · Cost engine

Cache hit rate

The share of runs that resolved to a compiled skill instead of a fresh compile — the single number that says whether the cost engine is working.

What is cache hit rate?

Cache hit rate is (hits + adapts) ÷ (hits + adapts + compiles) over a window of your runs. Twin computes it per tenant from the mode each run settled in, alongside total credits charged and an estimate of the credits those hits and adapts saved against cold compiles.

Why it matters

It is the metric that tells you whether your agents are re-phrasing the same intents (in which case the cache is compounding) or genuinely doing new things every time (in which case it is not, and that is worth knowing). It is a free read — it does not reserve or charge credits.

Related terms

See it in context

Back to the full glossary

Put your agent to work. Keep the decision.

Give your agent a real browser, authorize the accounts it may use, and keep control of every step. Repeated runs compile into skills that replay without a model.