[StevenWaterman]

Cached tokens are cheaper (90% discount ish) but not free

[moyix]

Also, unlike OpenAI, Anthropic's prompt caching is explicit (you set up to 4 cache "breakpoints"), meaning if you don't implement caching then you don't benefit from it.

[netcraft]

thats a very generous way of putting it. Anthropic's prompt caching is actively hostile and very difficult to implement properly.