---
id: 20260715-T0-05
title: "编码智能体需要多少上下文？研究发现大部分阅读都是浪费"
title_en: "Coding Agents Waste Most Context Reading; Study Reveals What They Actually Need"
url: https://ai.daily.yangsir.net/daily/20260715-T0-05
issue_date: 2026-07-15
publish_date: 2026-07-14T04:00:00.000Z
source_name: "arXiv cs.LG (ML)"
source_url: https://arxiv.org/abs/2607.09691
---

# 编码智能体需要多少上下文？研究发现大部分阅读都是浪费

arXiv 新论文指出，现代编码智能体虽能读取整个代码仓库，但绝大部分上下文被浪费。研究者聚焦于智能体在行动时真正需要的信息，而非其能力上限。该发现提示开发者在设计AI编码工具时，应更精准地筛选输入内容，而非追求更大的上下文窗口。

## English Version

**Coding Agents Waste Most Context Reading; Study Reveals What They Actually Need**

A new arXiv study finds modern coding agents waste most of their context reading entire repositories. The research focuses on what information agents actually need at the moment of action, not just capacity limits. This suggests developers should prioritize precise input filtering over larger context windows.

---

**来源**：[arXiv cs.LG (ML)](https://arxiv.org/abs/2607.09691)

**详情页**：https://ai.daily.yangsir.net/daily/20260715-T0-05

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*