待验证50% 置信事实时间未知
A deep programming session with an AI Agent may accumulate tens of thousands of tokens of state information, including code modification sequences, debugging hypotheses, and architectural reasoning chains
1
来源数
50%
置信度
中期 (~90 天)
时效性
2026/7/2
首次发现
有效期至：2026/9/30
来源
相关事实
待验证AI编程任务需将整个项目代码结构、依赖关系、错误信息送入模型，Token消耗是普通对话的数倍乃至数十倍77% 相似待验证A single conversation on a medium-to-large project can consume tens or even hundreds of thousands of Tokens when using AI coding assistants75% 相似待验证AI Agents powered by large language model reasoning can understand ambiguous instructions, process unstructured information, and make decisions in unforeseen situations.74% 相似待验证AI编程模型在完成任务时需要消化大量上下文，包括当前文件、相关代码库、项目结构和依赖关系，构成海量输入Token74% 相似已验证AI编码代理一次完整的任务执行可能涉及数万甚至数十万token的消耗73% 相似
引用此条事实
Stable URI
https://kongchang.com/claim/56156API
curl https://kongchang.com/api/v1/knowledge/claims/56156MCP
get_claim(id=56156)