{"version":"1.0","generated_at":"2026-10-25T10:56:23.909406","id":3598,"slug":"fancy-harnesses-make-coding-agents-more-expensive-but-not-better-study-finds","title":"研究发现：花哨的编码代理 Harness 更贵但不更好","summary":"山东大学两位研究者固定模型后比较 Claude Code、OpenCode 与 mini-SWE-agent，发现更换 harness 带来的结果差异大多落在同一设置重跑的噪声范围内，而运行成本可能增至三倍。唯一可测量的差异是 OpenCode 稳定落后，部分原因可追溯到输出限制处理。","abstract":"研究发现：花哨的编码代理 Harness 更贵但不更好 山东大学两位研究者固定模型后比较 Claude Code、OpenCode 与 mini-SWE-agent，发现更换 harness 带来的结果差异大多落在同一设置重跑的噪声范围内，而运行成本可能增至三倍。唯一可测量的差异是 OpenCode 稳定落后，部分原因可追溯到输出限制处理。 山东大学两位研究者固定模型后比较 Claude Code、OpenCode 与 mini-SWE-agent，发现更换 harness 带来的结果差异大多落在同一设置重跑的噪声范围内，而运行成本可能增至三倍。唯一可测量的差异是 OpenCode 稳定落后，部分原因可追溯到输出限制处理。 原贴提到：Makers tout score gains from new agent harnesses like Claude Code or Ope 来源：the-decoder.com 制造商吹捧像 Claude Code 或 OpenCode 这样的新 agent harness 带来分数提升。一项研究发现，这种说法站不住脚。更换 harness 对结果的影响，几乎不比把同一套设置跑两遍更大，但成本可能增至三倍。一个编码代理将语言模型与 harness 配对，harness 是系统提示、工具和上下文管理，使模型能在代码仓库中工作。每天都有新 harness 发布，排行榜随意混用它们，但山东大学的两位研究人员测量了在保持模型不变时，harness 实际有多重要。论文标题直接给出答案：“What does a harness buy? Mostly…","access_level":"member","access_label":"免费阅读","access_mode":"full","is_preview":false,"canonical_url":"https://opc.beizhux.com/content/3598/fancy-harnesses-make-coding-agents-more-expensive-but-not-better-study-finds","html_url":"https://opc.beizhux.com/content/3598/fancy-harnesses-make-coding-agents-more-expensive-but-not-better-study-finds","json_url":"https://opc.beizhux.com/content/3598/fancy-harnesses-make-coding-agents-more-expensive-but-not-better-study-finds.json","published_at":"2026-10-24T17:21:39","updated_at":"2026-10-25T10:04:17","category":{"slug":"ai-business","name":"AI小生意项目库"},"source":{"site":"the-decoder.com","author":"Jonathan Kemper","url":"https://the-decoder.com/fancy-harnesses-make-coding-agents-more-expensive-but-not-better-study-finds/"},"tags":["AI编程","Harness","SWE-bench","基准测试","开发者工具","成本效率","编码代理","编码智能体","评测方法"],"topics":[{"slug":"agent-workflow","name":"Agent工作流","url":"https://opc.beizhux.com/topics/agent-workflow","reason":"这篇内容命中「Agent、智能体、Claude Code」等主题信号。"},{"slug":"ai-side-business","name":"AI副业","url":"https://opc.beizhux.com/topics/ai-side-business","reason":"这篇内容命中「项目、小生意、变现」等主题信号。"},{"slug":"ai-tools","name":"AI工具","url":"https://opc.beizhux.com/topics/ai-tools","reason":"这篇内容命中「工具、模型、Claude」等主题信号。"}],"keywords":["AI小生意项目库","Agent工作流","AI副业","AI工具","工具","自动化","模型","Cursor","Claude","Agent","智能体","工作流"],"questions":[{"question":"研究发现：花哨的编码代理 Harness 更贵但不更好主要讲什么？","answer":"山东大学两位研究者固定模型后比较 Claude Code、OpenCode 与 mini-SWE-agent，发现更换 harness 带来的结果差异大多落在同一设置重跑的噪声范围内，而运行成本可能增至三倍。唯一可测量的差异是 OpenCode 稳定落后，部分原因可追溯到输出限制处理。"},{"question":"这篇文章最值得关注的要点是什么？","answer":"山东大学两位研究者固定模型后比较 Claude Code、OpenCode 与 mini-SWE-agent，发现更换 harness 带来的结果差异大多落在同一设置重跑的噪声范围内，而运行成本可能增至三倍。唯一可测量的差异是 OpenC…；原贴提到：Makers tout score gains from new agent harnesses like Claude Code or Ope；来源：the-decoder.com"},{"question":"这篇文章和哪些AI专题相关？","answer":"它适合放在Agent工作流、AI副业、AI工具专题里阅读。 关联原因：这篇内容命中「Agent、智能体、Claude Code」等主题信号。；这篇内容命中「项目、小生意、变现」等主题信号。；这篇内容命中「工具、模型、Claude」等主题信号。"},{"question":"阅读这篇文章建议先理解哪些关键词？","answer":"建议先理解AI工具、工具、自动化、模型、Cursor这些关键词，再结合正文判断工具、机会或风险是否值得进入自己的工作流。"}],"terms":[{"slug":"ai-tools-term","name":"AI工具","definition":"在AI觉醒星球里，「AI工具」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/ai-tools-term","path":"/glossary/ai-tools-term","json_url":"https://opc.beizhux.com/glossary/ai-tools-term.json"},{"slug":"tools","name":"工具","definition":"在AI觉醒星球里，「工具」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/tools","path":"/glossary/tools","json_url":"https://opc.beizhux.com/glossary/tools.json"},{"slug":"automation","name":"自动化","definition":"在AI觉醒星球里，「自动化」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/automation","path":"/glossary/automation","json_url":"https://opc.beizhux.com/glossary/automation.json"},{"slug":"ai-model","name":"模型","definition":"在AI觉醒星球里，「模型」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/ai-model","path":"/glossary/ai-model","json_url":"https://opc.beizhux.com/glossary/ai-model.json"},{"slug":"cursor","name":"Cursor","definition":"在AI觉醒星球里，「Cursor」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/cursor","path":"/glossary/cursor","json_url":"https://opc.beizhux.com/glossary/cursor.json"},{"slug":"claude","name":"Claude","definition":"在AI觉醒星球里，「Claude」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/claude","path":"/glossary/claude","json_url":"https://opc.beizhux.com/glossary/claude.json"},{"slug":"agent","name":"Agent","definition":"在AI觉醒星球里，「Agent」属于「Agent工作流」方向。聚合Agent、AI编程、Claude Code、Cursor、工作流自动化等内容，帮助你把智能体能力落到真实任务链路。 把智能体当成任务系统，而不是聊天窗口。","topic_slug":"agent-workflow","topic_name":"Agent工作流","topic_title":"Agent工作流：智能体、AI编程与自动化协作","topic_url":"https://opc.beizhux.com/topics/agent-workflow","topic_path":"/topics/agent-workflow","url":"https://opc.beizhux.com/glossary/agent","path":"/glossary/agent","json_url":"https://opc.beizhux.com/glossary/agent.json"},{"slug":"ai-agent","name":"智能体","definition":"在AI觉醒星球里，「智能体」属于「Agent工作流」方向。聚合Agent、AI编程、Claude Code、Cursor、工作流自动化等内容，帮助你把智能体能力落到真实任务链路。 把智能体当成任务系统，而不是聊天窗口。","topic_slug":"agent-workflow","topic_name":"Agent工作流","topic_title":"Agent工作流：智能体、AI编程与自动化协作","topic_url":"https://opc.beizhux.com/topics/agent-workflow","topic_path":"/topics/agent-workflow","url":"https://opc.beizhux.com/glossary/ai-agent","path":"/glossary/ai-agent","json_url":"https://opc.beizhux.com/glossary/ai-agent.json"}],"preview_hidden_sections":[],"related_items":[{"id":3601,"title":"Krea 开放 Krea 3 图像编辑 beta，支持 2k 快速生成与多轮精确编辑","url":"https://opc.beizhux.com/content/3601/krea-krea-3-beta-2k","summary":"Krea 上线 Krea 3 图像编辑 beta，端到端 2k 编辑 8 秒内完成，画质对标 Nano Banana Pro 与 GPT Image 2.5，支持 bounding box 控制与精确多轮编辑，官方演示展示了局部快速改图效果。","access_level":"member"},{"id":3604,"title":"AI 墓地：一份未能活下来的项目和创业公司清单","url":"https://opc.beizhux.com/content/3604/the-ai-graveyard-a-running-list-of-projects-and-startups-that-didn-t-make-it","summary":"TechCrunch AI 汇总 AI 项目失败清单：Relay 关停、Character.ai 破产，S&P 称约42%企业AI计划被放弃；OpenAI 将独立功能并入 ChatGPT，Apple 新 Siri 延期并引发2.5亿美元和解。","access_level":"member"},{"id":3605,"title":"Block 发布新语音模型 Dessa：本地运行，听起来「像人」到令人不安","url":"https://opc.beizhux.com/content/3605/block-s-new-voice-model-sounds-eerily-human","summary":"Jack Dorsey 旗下的 Block 研究实验室发布全双工语音模型 Dessa，完全以人类语音训练，可在本地与消费级 GPU 上运行，自带笑声、停顿、哭腔等情绪表达，主张把「情感」直接内建于模型，而非依赖文本管线或语音标签驱动。","access_level":"member"}],"usage_policy":{"summarizable":true,"preferred_url":"https://opc.beizhux.com/content/3598/fancy-harnesses-make-coding-agents-more-expensive-but-not-better-study-finds","private_fields_excluded":true}}