Unverified50% confidenceFactExact time
Cached input refers to the model caching and reusing repeated or previously processed content, with billing rates typically 10%-25% of the normal input price.
1
Sources
50%
Confidence
Long-term
Relevance
10/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedCache hit discounts target repeated system prompts or context prefixes: if this content has already been cached, the provider does not need to recompute it and can heavily discount the cost of the hit portion, commonly to 10% to 25% of the original price.75% similarUnverifiedCached input tokens typically enjoy a discount of 25%–50% off the original price, as there is no need to recompute embeddings and attention weights—only the precomputed KV vectors need to be loaded from cache.74% similarUnverifiedPrompt cache hit determination is based on an exact match of "prefix + cache key", where the cache key includes some request metadata; cache writes are typically billed at 1.25 times the base input price.72% similarUnverifiedAtmoRouter charges cached input at 10% of the input price.71% 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.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1059472API
curl https://kongchang.com/api/v1/knowledge/claims/1059472MCP
get_claim(id=1059472)