Verified65% confidenceFactExact time
Each model response includes the full conversation history (context window) in billing, and as the number of dialogue turns increases, the token consumption per call grows linearly or even super-linearly.
3
Sources
65%
Confidence
Long-term
Relevance
7/22/2026
First Seen
Sources
Related Claims
VerifiedIn multi-turn conversations, before each new response is generated, the system must re-input all prior conversation history into the model, causing token consumption to accumulate and grow with each turn83% similarUnverifiedDue to the cyclic reasoning nature of Agent systems, the complete conversation history is resent to the model after each tool call, causing Token consumption to grow super-linearly.80% similarUnverifiedThe cost of each large language model call is positively correlated with the number of input tokens, and in an Agent loop, as the conversation history accumulates, the cost of each call grows linearly.79% similarUnverifiedMulti-turn dialogue is implemented by re-inputting all previous conversation history to the model with each request, rather than the model having genuine memory capability.76% similarUnverified模型在每次推理时需要将当前对话的完整历史记录一并输入处理，总结全部内容操作会将整个上下文历史计入输入费用75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/584993API
curl https://kongchang.com/api/v1/knowledge/claims/584993MCP
get_claim(id=584993)