Unverified50% confidenceFactExact time
Both Anthropic's Claude API and OpenAI's API natively support prompt caching, with input token prices on cache hits typically only 10%–25% of the normal price.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/8/2026
First Seen
Valid until: 11/6/2026
Sources
Databricks如何将AI编码成本砍掉70%：策略与启示
hackernewshackernews8/7/2026
Related Claims
VerifiedOpenAI's API billing distinguishes between input tokens and output tokens, with output tokens typically priced higher.79% similarUnverifiedOpenRouter's billing method charges by token usage, and prices are generally on par with or slightly higher than the official API.77% similarUnverifiedOpenAI has introduced a prompt cache breakpoint mechanism that allows developers to manually specify cache breakpoints, with cache-hit portions typically billed at a discounted rate of 10%–25% of the original price.75% similarUnverifiedThe API cache hit mechanism allows input tokens that have already been sent to enjoy discounted pricing on repeated calls, typically at 10%–50% of the original price.74% similarUnverifiedThe prompt caching feature introduced by Anthropic and OpenAI operates on the server side, offering a 50–90% price discount for content with identical long prefixes.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/711964API
curl https://kongchang.com/api/v1/knowledge/claims/711964MCP
get_claim(id=711964)