DeepAnima
Kouri Ai

Cache Pricing

How cache billing works for models

You can tell whether a model supports caching by checking if it has cache pricing in the model price list.

Read-Only Price

If a model has read price only: caching is implicit and requires no action. When a request hits the cache, you are charged the read cache price for the cached amount.

Read Price + Write Price

If a model has both read price and write price: you must first perform a write according to the official documentation of the corresponding endpoint; you are charged the write price. If a later request hits the cache, you are charged the read cache price for the cached amount.

Notes

Neither approach guarantees a cache hit rate. Use models that require writing to cache with caution.

On this page