Kimi API adds 5-minute and 1-hour context-cache TTLs
Kimi says cached content for kimi-k3 is billed at one-tenth the cache-miss price. Cache Write is billed separately, with 5-minute and 1-hour TTL options; omitting the TTL setting defaults to 5 minutes.
SessionWatcher editorial · Published · Updated · Source announcement: 2026-09-28
Related guide: How to check your Kimi usage
Kimi explains its context-cache pricing
Kimi says context caching reuses repeated request prefixes, including codebases and tool definitions. For kimi-k3, cache hits cost one-tenth of cache misses.
Cache Write is a separate billed item. Kimi offers 5-minute and 1-hour time-to-live options, and says requests that omit prompt_cache_options use the 5-minute TTL by default. Cache Write charges apply when prefixes that meet the hit conditions are written automatically.
Sources: Context caching upgrade: 5 min & 1 hour TTL options
Choosing a cache lifetime
Kimi recommends 5 minutes when follow-up requests usually arrive within that period. Its 1-hour option is intended for prefixes that may be reused more than five minutes later but within an hour. It advises against configuring caching for prefixes typically reused only after more than an hour.
The announcement also says cached prefixes are isolated by organization, cannot be cleared manually, and expire after being inactive for their selected TTL.
Sources: Context caching upgrade: 5 min & 1 hour TTL options
AI assisted reporting, checked against the linked official sources. Source pages checked 2026-10-03. Editorial process and corrections.