Prompt Cache Hit Planner — Estimate LLM Cache Savings

Pricing setup

Provider preset Base input price ($ / 1M tokens)

Cache write multiplier

Cache read (hit) multiplier

Write multiplier: cost to write into cache (usually >1x). Read multiplier: cost of a cache hit (usually much cheaper, e.g. 0.1x).

Shared prefix tokens (system/context) Unique tokens per request (not cached)

Avg turns per session

Requests per day

First request in each session is assumed to be a cache write (miss); remaining turns hit the cache if within TTL.

0%

estimated cache hit rate

Cost without cache / day $0

Cost with cache / day $0

Savings / day $0

Savings / month $0

Savings % 0%