Use case
Claude Opus 4.8 prompt caching
prompt caching notes for claude-opus-4-8: 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 $2.00 on corouter (input $1.00, output $1.00).
Cache-read rate
Cache read for claude-opus-4-8 lists at $0.500 and is $0.100 on corouter per million tokens. Uncached input is $1.00.
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 Opus 4.8 matches the Opus 5 list rate and window, without the Opus 5.5 cache-read cut. Send claude-opus-4-8 to keep an Opus 4.8 client unchanged.
Questions
- Which model ID should the request send?
- Send claude-opus-4-8. 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.