Unverified50% confidenceFactExact time
Active Memory abstracts historical conversations into structured facts and retrieves and injects only a small number of relevant vectors, with cache reads costing only $0.20 per million tokens, far lower than the cost of repeatedly feeding in historical conversations.
1
Sources
50%
Confidence
Long-term
Relevance
10/13/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAccording to the author's logs, actual input content accounted for only about 0.01% of the bill, with 96% of the cost coming from old conversation history being resent repeatedly.70% similarUnverifiedFully automated memory governance remains immature at present, and low-cost human intervention entry points should be preserved.65% similarVerifiedThe cache read mechanism retains the KV cache for recurring long prompt prefixes, charging at an extremely low rate upon a cache hit; the leaked cache read price for Opus 5.5 is approximately 1/20 of the input price.63% similarUnverifiedMemo is free when saving fewer than ten thousand times and searching fewer than one thousand times per month, and the cost of running a local model can be zero.62% similarUnverifiedOn most Claude models, Anthropic bills cache reads at 0.1× the base input price; default 5-minute cache writes at 1.25×; 1-hour cache writes at 2×; and a 5-minute cache entry is refreshed for free each time it is used.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1020387API
curl https://kongchang.com/api/v1/knowledge/claims/1020387MCP
get_claim(id=1020387)