Use case
Claude Haiku 4.5 prompt caching
prompt caching notes for claude-haiku-4-5: the rate card, a worked volume, and when to send this ID. A sample of 1 million input tokens plus 200 thousand output tokens is about $0.400 on corouter (input $0.200, output $0.200).
Cache-read rate
Cache read for claude-haiku-4-5 lists at $0.100 and is $0.020 on corouter per million tokens. Uncached input is $0.200.
Only the input that actually hits cache is billed at the cache-read rate. This page does not promise your prompt will hit.
Where this ID sits
Claude Haiku 4.5 is the short-window, low list-rate Claude row: 200K, chat and vision, no Code tag. Send claude-haiku-4-5 for volume that fits that window.
Questions
- Which model ID should the request send?
- Send claude-haiku-4-5. Do not put a different name in the model field.
- What is the minimum top-up?
- Prepaid credits, $5 minimum top-up, no subscription tier.