Unverified50% confidenceFactExact time
多轮对话中第N轮的输入成本是前N-1轮内容的总和，累积效应会导致成本抛物线式增长
1
Sources
50%
Confidence
Long-term
Relevance
8/18/2026
First Seen
Sources
Related Claims
Unverified智能体每轮调用均需传入完整上下文导致输入Token随对话轮次膨胀，被称为「上下文窗口通胀」，是成本失控的主要根源75% similarUnverified随着对话轮次增加，每轮需携带完整历史上下文，输入token数量持续膨胀，导致推理时间呈现非线性增长73% similarUnverifiedAgent间通信会导致上下文窗口线性增长使成本呈指数级攀升，以及长对话链中的「目标漂移」现象73% similarUnverifiedAs multi-turn tool calls and conversations accumulate in Agent tasks, context expansion causes inference costs to rise exponentially and response latency to increase71% similarUnverified向量维度越高语义表达能力越强，但存储开销和检索延迟也随之增加70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/767505API
curl https://kongchang.com/api/v1/knowledge/claims/767505MCP
get_claim(id=767505)