Unclear It is not clear where the time limit comes from or whether the wait is switched on for users.
What
Claude Code sometimes sends a request in advance to warm the prompt cache, a store that lets repeated text be handled faster. A turn can now wait for such a request to finish, up to a time limit. The warm request is cancelled if it used a different model or started too recently.
Why
This affects how quickly Claude starts replying when a cache-warming request is still running.
Something disagreesSomething we can check disagrees with this entry, or the writer said they could not settle it.
The writer flagged doubt
It is not clear where the time limit comes from or whether the wait is switched on for users.