三份 Agents.md 实测对比：耗时、token 用量与成本差异明显
有人拿日志实测了三种 Agents.md 的生成成本，Opus 5.5 那份只要 15.5M tokens、约 $0.25，价格敏感的话可以看看数字再决定。
一次基于日志审计的对比测试给出了三组数据：手写的 Agents.md 跑了约 42 分钟、消耗 49M tokens、成本约 $0.
48；DeepSeek 在 DSH 中生成的 Agents.
md 用时约 35 分钟、42M tokens、约 $0.33；
Opus 5.5 在 CC 中生成的版本最快，约 25 分钟、仅 15.
5M tokens、约 $0.25，但生成的方案更简单。
三组结果均来自实际运行日志而非自述数据。
Tests based on the audit of the logs: QT) my old handwrittenaAgents.md. ≈42 min, 49M tokens, ≈$0,48 A) Agents.md by DeepSeek in DSH. ≈35 min, 42M tokens, ≈$0.33 B) Agents.md by Opus 5.5 in CC. ≈25 min, only 15.5M tokens, ≈$0.25. but simpler You can do better here I think https://t.co/hgvXRJIpha
This report is categorized under AITOP's "How-to" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Agents.md, Opus 5.5, DeepSeek, 提示词工程.
Source attribution: curated from Teortaxes (@teortaxesTex). Cross-source corroboration: 2 independent sources have covered this development (including shao__meng), confirming industry-wide signal strength.
- shao__meng10-17 01:04Source
