Prompt caching

Prompt Caching Savings Calculator

Compare uncached and cached input costs for repeated prompts, long system instructions, shared RAG context, and high-volume AI workflows.

Estimate prompt caching savings

Model

Requests per month

Cache hit rate percent

Reusable input tokens

Dynamic input tokens

Output tokens per request

Estimated monthly savings$0.00

Without caching$0.00

With caching$0.00

Savings rate0%

Cached input price$0.00

How caching changes the bill

Caching only helps when a meaningful part of the input is reused. The estimate comparesstandard input cost with cached reusable input + uncached dynamic input + output cost.

Output tokens usually do not become cheaper through input caching.
Provider cache rules can depend on model, minimum prefix length, and cache duration.
Higher hit rates make caching more valuable.