Claude usage
Loading…
Plan limits
Total tokens processed
–
Tokens per day
Stacked by model
Where the tokens go
Most input is re-read from the prompt cache, which is far cheaper than fresh input.
Models
Share of API responses
When I use Claude
Responses by weekday and hour