Anthropic says prompt caching can cut repeated-context token costs by up to 90%

aitrendz_xyz · x · 2026-07-27

Prompt caching lets API builders avoid paying repeatedly for large, repeated context. Cached content is stored server-side and reused, which can make cached tokens up to 90% cheaper and faster. The cache lasts 5 minutes, refreshes on each use, and only works once a block crosses the minimum size threshold.

Original post →

More from coding & agent

coding & agent channel →