Unverified50% confidenceFactExact time
When an LLM streaming request is preempted and the entire prompt is resent, you pay again for the already-generated prefix tokens.
1
Sources
50%
Confidence
Long-term
Relevance
10/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAnthropic's Prompt Caching feature grants a discount on the corresponding tokens if the fixed prefix portion of a request can reuse the cache.69% similarUnverifiedBatch processing bundles large numbers of requests for submission, processes them asynchronously on the server side, and returns results collectively, usually accompanied by more favorable pricing.65% similarUnverifiedEach retry resends the complete prompt input tokens and is billed in full; a request with 4,000 input tokens retried 3 times will consume 16,000 tokens.65% similarUnverifiedThe Prompt Caching feature introduced by Anthropic in 2025 offers a 90% price discount for repeatedly occurring prefix contexts, but only applies to consecutive requests within the same session or from the same user.63% similarUnverifiedIn the demo, after entering the customer message "I was charged twice, please refund the duplicate payment," the model returned the billing category with a probability of approximately 98%.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/999884API
curl https://kongchang.com/api/v1/knowledge/claims/999884MCP
get_claim(id=999884)