Unverified50% confidenceFactExact time
RadixAttention can identify prefix segments shared across different requests and directly reuse cached computation results, reducing redundant computation in scenarios with fixed system prompts, multi-turn conversations, or batch document Q&A, thereby improving throughput and reducing time-to-first-token (TTFT).
1
Sources
50%
Confidence
Long-term
Relevance
9/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedTypical scenarios that benefit from prefix caching include multi-turn conversations sharing the same system prompt, repeated document contexts in RAG, and stable codebase prefixes in code completion.74% similarUnverifiedSGLang's RadixAttention technology can automatically reuse KV Cache across multiple requests, managing shared prefixes through a radix tree structure to reduce redundant computation.73% similarUnverified语义缓存（Semantic Cache）通过计算新查询与历史查询的向量相似度，超过阈值时直接返回缓存答案，GPTCache等开源工具实现了这一机制，通常与TTL机制结合66% similarUnverifiedAnthropic's prompt caching feature allows caching of repeatedly occurring system prompts or code files to reduce input token costs.65% similarUnverifiedColBERT 采用延迟交互机制，为查询和文档的每个 token 生成嵌入向量，通过 MaxSim 操作计算相似度64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/931677API
curl https://kongchang.com/api/v1/knowledge/claims/931677MCP
get_claim(id=931677)