技巧

用 Agent Skills 让 Claude Haiku 替代 Opus 5.5,快 7 倍省一半 token

精选理由

教你不改模型就能让 Haiku 干 Opus 5.5 的活:59 秒替代 7 分钟,token 省一半,思路挺实用。

作者用 Zep Agent Skills 教 Claude Haiku 完成原本由 Opus 5.5 承担的任务,方式类似微调但不需要改动模型本身。效果上,同一任务 Opus 5.5 耗时 7 分钟、消耗 140 万 token,Haiku 只需 59 秒和 70 万 token,速度提升 7 倍、token 消耗减半。文章思路是把大模型的任务经验封装成技能,交给小模型执行以降低成本。

原文 · Ate-a-Pi

How to replace a large, expensive, and slow model with a small, cheaper, fast model and still get the same results.

7x faster and 50% fewer tokens.

This is like fine-tuning, but without touching the model.

Here, we teach Haiku, using Zep Agent Skills, to do what Opus 5.5 does to complete a task.

• It takes 7 minutes with Opus 5.5. Only 59 seconds with Haiku. • It takes 1.4M tokens with Opus 5.5. Only 700K with Haiku.