It was 9.4B to be exact and it was over the course of two days lol. It was between two projects so 99% of them were cached reads.

The GPT was about 1B on two projects on 300$ worth of plans all on Astra and I capped out on usage.

Anthropic caching must be better because the cache rates are better on Claude models.