Unverified50% confidenceFactExact time
In the agent's multiple rounds of iteration, the context window keeps expanding with each iteration, causing the number of input tokens per inference round to increase, and the cost of attention computation grows superlinearly.
1
Sources
50%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedWhen an agent is in a loop, each call carries the accumulated conversation history, causing token consumption to grow linearly or even super-linearly with the number of iterations.77% similarUnverified多Agent架构中Agent间传递上下文时重复信息被反复处理，导致Token消耗和成本上升76% similarUnverified在多轮迭代模式下每一轮执行都会重新传入累积的上下文，导致token消耗呈近似二次方增长73% similarUnverifiedAgents burn through more Tokens than ordinary chat because executing tasks requires multiple think-act-observe loops, and intermediate results are fed back into the model, causing the context window to expand.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1030786API
curl https://kongchang.com/api/v1/knowledge/claims/1030786MCP
get_claim(id=1030786)