[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-codex-third-party-model-integration-guide-en":3,"article-related-codex-third-party-model-integration-guide-en":31,"series-ai-agent-07fb3bcc-9f38-4153-a9c8-5d67ba7f5018":80},{"id":4,"slug":5,"title":6,"content":7,"summary":8,"source":9,"source_url":10,"author":11,"image_url":12,"cover_image":12,"category":13,"language":14,"translated_content":11,"related_article_id":15,"keywords":16,"key_takeaways":23,"views":27,"created_at":28,"published_at":29,"topic_cluster_id":30},"07fb3bcc-9f38-4153-a9c8-5d67ba7f5018","codex-third-party-model-integration-guide-en","Codex 接入第三方模型完整指南","\u003Cp data-speakable=\"summary\">\u003Ca href=\"\u002Ftag\u002Fopenai\">OpenAI\u003C\u002Fa> \u003Ca href=\"\u002Ftag\u002Fcodex\">Codex\u003C\u002Fa> App、CLI 和 SDK 现可接入第三方开源模型。\u003C\u002Fp>\u003Cp>这篇指南适合想把 OpenAI \u003Ca href=\"https:\u002F\u002Fplatform.openai.com\u002Fdocs\" target=\"_blank\" rel=\"noopener noreferrer\">Codex 文档\u003C\u002Fa>和 \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fopenai\u002Fcodex\" target=\"_blank\" rel=\"noopener noreferrer\">Codex GitHub 仓库\u003C\u002Fa>接到自定义模型上的开发者。跟着做完，你会得到一套可运行的接入流程，能把第三方模型挂到 Codex App、CLI 或 SDK 上，并验证它是否真的在工作。\u003C\u002Fp>\u003Cp>本文只基于已公开的信息做落地整理，重点放在准备环境、配置模型提供方、连通 Codex 工具链、以及最后的验证和排错。你不需要先改造整个项目，只要先把模型端和 Codex 端对接起来，就能开始试用。\u003C\u002Fp>\u003Ch2>Before you start\u003C\u002Fh2>\u003Cul>\u003Cli>OpenAI 账号，且能访问 Codex 相关功能。\u003C\u002Fli>\u003Cli>一个可用的第三方模型提供方账号，支持 OpenAI 兼容 API 或 Codex 兼容接入方式。\u003C\u002Fli>\u003Cli>API key：OpenAI Codex 侧 key，以及第三方模型侧 key。\u003C\u002Fli>\u003Cli>Node.js 20+，如果你要用 CLI 或 SDK 的 JavaScript 示例。\u003C\u002Fli>\u003Cli>Git 2.40+，用于拉取示例项目或仓库。\u003C\u002Fli>\u003Cli>至少 2 GB 可用内存，便于本地运行 CLI、日志和测试脚本。\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>Step 1: 获取 Codex 工具链\u003C\u002Fh2>\u003Cp>目标是先把 Codex 的 App、CLI 或 SDK 入口准备好，这样后续才能切换模型而不是停留在文档阶段。先确认你打算从哪个入口开始：如果偏向命令行，就装 CLI；如果偏向代码集成，就先接 SDK；如果想快速试验，就用 App。\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782396176284-xvyq.png\" alt=\"Codex 接入第三方模型完整指南\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cpre>\u003Ccode>git clone https:\u002F\u002Fgithub.com\u002Fopenai\u002Fcodex.git\ncd codex\n# 按仓库说明安装 CLI 或 SDK 依赖\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>验证方式很简单：你应该能看到仓库内容、安装命令成功执行，并且本地能打开 Codex 的配置说明或帮助信息。\u003C\u002Fp>\u003Ch2>Step 2: 准备第三方模型端点\u003C\u002Fh2>\u003Cp>目标是拿到一个可调用的模型 \u003Ca href=\"\u002Ftag\u002Fapi\">API\u003C\u002Fa> 地址、模型名和密钥。Codex 官方说法是可以搭配任何开源模型使用，所以关键不是“模型是不是 OpenAI 的”，而是它是否能提供稳定的推理接口。优先选择支持 OpenAI 风格请求格式的服务，这样迁移成本最低。\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782396180532-1qwt.png\" alt=\"Codex 接入第三方模型完整指南\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>把这些信息记下来：base URL、model id、API key、请求限流规则，以及是否需要额外的组织 ID 或项目 ID。后面配置时，这些字段会直接映射到 Codex 的连接参数。\u003C\u002Fp>\u003Cp>验证方式是发一次最小请求。你应该能拿到一个正常的文本回复，而不是 401、404 或格式错误。\u003C\u002Fp>\u003Ch2>Step 3: 配置 Codex 模型连接\u003C\u002Fh2>\u003Cp>目标是让 Codex 读到你的第三方模型配置，并把请求转发过去。不同入口的配置方式会不同，但思路一致：把模型提供方的地址、密钥和模型名写进 Codex 的配置文件或环境变量。\u003C\u002Fp>\u003Cpre>\u003Ccode>export OPENAI_API_KEY=\"your-codex-key\"\nexport OPENAI_BASE_URL=\"https:\u002F\u002Fyour-third-party-endpoint\u002Fv1\"\nexport OPENAI_MODEL=\"your-model-name\"\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>如果你用的是支持自定义 provider 的配置文件，就把同样的信息写入对应字段。重点是确认 Codex 发出的请求不再落到默认模型上，而是命中你填的第三方端点。\u003C\u002Fp>\u003Cp>验证方式是启动 Codex CLI 或 SDK 示例后查看日志。你应该能看到请求发往第三方 base URL，并且返回内容来自你选择的模型。\u003C\u002Fp>\u003Ch2>Step 4: 运行一次端到端任务\u003C\u002Fh2>\u003Cp>目标是证明接入不是“能连上”，而是“能干活”。选一个小任务最合适，比如让 Codex 生成一个函数、修改一个测试、或总结一段代码。任务越小，越容易判断模型是否真正被调用。\u003C\u002Fp>\u003Cp>建议先用 CLI 跑一个单文件任务，再用 SDK 跑一个简短脚本。这样你能分别验证交互式流程和程序化流程，避免只在一个入口上成功。\u003C\u002Fp>\u003Cp>验证方式是你应该看到模型返回可执行的代码、补丁或解释，并且输出结果符合你指定的第三方模型风格和能力边界。\u003C\u002Fp>\u003Ch2>Step 5: 固化环境变量和回退方案\u003C\u002Fh2>\u003Cp>目标是把临时接入变成可重复使用的配置。把 base URL、模型名和 key 放进 `.env`、CI secrets 或部署环境变量里，避免每次手动输入。与此同时，准备一个回退方案，方便第三方模型不可用时切换回默认提供方。\u003C\u002Fp>\u003Cp>如果你的团队要多人协作，最好把配置写进 README 或内部运行手册，并注明哪些字段必须保密，哪些字段可以公开。这样后续换模型时，改动范围会更小。\u003C\u002Fp>\u003Cp>验证方式是重启终端或重新拉起容器后，Codex 仍然能自动读到配置，不需要人工补参。\u003C\u002Fp>\u003Ch2>Common mistakes\u003C\u002Fh2>\u003Cul>\u003Cli>把模型名写错。修复方法：先用第三方控制台的模型列表复制准确名称，再写入配置。\u003C\u002Fli>\u003Cli>端点不是 OpenAI 兼容格式。修复方法：确认路径、请求体和认证头与 Codex 期望一致，必要时加一层适配代理。\u003C\u002Fli>\u003Cli>只测了连通性，没测任务输出。修复方法：跑一次真实代码生成或修改任务，检查返回结果是否可用。\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>What's next\u003C\u002Fh2>\u003Cp>接下来可以把同一套接入方式扩展到团队工作流里，比如接入 CI、自动化代码审查，或者为不同任务路由不同模型；如果你需要更稳的方案，再继续研究 Codex 的配置层、日志层和模型回退策略。\u003C\u002Fp>","OpenAI Codex App、CLI 和 SDK 现可接入第三方开源模型。","zhuanlan.zhihu.com","https:\u002F\u002Fzhuanlan.zhihu.com\u002Fp\u002F2051149872900973800",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782396176284-xvyq.png","ai-agent","en","daf7edd7-d904-4f35-afd6-9caeac32c633",[17,18,19,20,21,22],"Codex","OpenAI","third-party models","CLI","SDK","OpenAI-compatible API",[24,25,26],"Codex App、CLI 和 SDK 可以接入第三方开源模型。","接入重点在于模型端点、API key 和模型名的正确映射。","先做最小请求和端到端任务验证，再固化到环境变量和回退方案。",0,"2026-06-25T14:02:29.820439+00:00","2026-06-25T14:02:29.807+00:00","a9bee732-b07c-4e5b-a0e6-3048577e32a7",{"tags":32,"relatedLang":39,"relatedPosts":43},[33,35,37],{"name":18,"slug":34},"openai",{"name":17,"slug":36},"codex",{"name":20,"slug":38},"cli",{"id":15,"slug":40,"title":41,"language":42},"codex-third-party-model-integration-guide-zh","Codex 接入第三方模型實作指南","zh",[44,50,56,62,68,74],{"id":45,"slug":46,"title":47,"cover_image":48,"image_url":48,"created_at":49,"category":13},"e2049534-8d94-453a-8cee-8eced0b74e69","public-sentry-keys-hijack-claude-code-cursor-en","Public Sentry keys can hijack Claude Code and Cursor","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782413277883-xtvj.png","2026-06-25T18:47:31.313932+00:00",{"id":51,"slug":52,"title":53,"cover_image":54,"image_url":54,"created_at":55,"category":13},"daccbfdf-46f3-432e-8b8d-aecb8198d1c1","loop-engineering-agent-completes-tasks-en","Loop Engineering 让 Agent 把事做完","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782408826341-e6s0.png","2026-06-25T17:33:18.472838+00:00",{"id":57,"slug":58,"title":59,"cover_image":60,"image_url":60,"created_at":61,"category":13},"0003f204-e4d0-4015-8208-bbd23ecfb908","grok-build-goal-autonomous-coding-en","Grok Build Adds \u002Fgoal for Autonomous Coding","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782374583821-11kl.png","2026-06-25T08:02:38.973865+00:00",{"id":63,"slug":64,"title":65,"cover_image":66,"image_url":66,"created_at":67,"category":13},"c41b19d2-48c8-4d88-92f9-d92d73cf9e90","set-up-ai-agent-workflows-5-practical-steps-en","Set Up AI Agent Workflows in 5 Practical Steps","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782314280123-7fv1.png","2026-06-24T15:17:28.642801+00:00",{"id":69,"slug":70,"title":71,"cover_image":72,"image_url":72,"created_at":73,"category":13},"61c1e05c-ea78-4f0a-b389-3f09eeabf7e3","anthropic-claude-tag-research-slack-search-en","Anthropic’s Claude Tag Research turns Slack into search","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782285508413-kwit.png","2026-06-24T07:18:03.3764+00:00",{"id":75,"slug":76,"title":77,"cover_image":78,"image_url":78,"created_at":79,"category":13},"8dbcd7ac-bae7-46c1-ba11-bdca1fd774e8","benchmark-harness-quality-beats-model-hype-coding-en","This benchmark proves harness quality beats model hype in coding","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782253062750-lxuw.png","2026-06-23T22:17:21.750686+00:00",[81,86,91,96,101,106,111,116,121,126],{"id":82,"slug":83,"title":84,"created_at":85},"03db8de8-8dc2-4ac1-9cf7-898782efbb1f","anthropic-claude-ai-agent-task-automation-en","Anthropic's Claude AI Agent: A New Era of Task Automation","2026-03-25T16:25:06.513026+00:00",{"id":87,"slug":88,"title":89,"created_at":90},"045d1abc-190d-4594-8c95-91e2a26f0c5a","googles-2026-ai-agent-report-decoded-en","Google’s 2026 AI Agent Report, Decoded","2026-03-26T11:15:23.046616+00:00",{"id":92,"slug":93,"title":94,"created_at":95},"e64aba21-254b-4f93-aa21-837484bb52ec","kimi-k25-review-stronger-still-not-legend-en","Kimi K2.5 review: stronger, still not a legend","2026-03-27T07:15:55.385951+00:00",{"id":97,"slug":98,"title":99,"created_at":100},"30dfb781-a1b2-4add-aebe-b3df40247c37","claude-code-controls-mac-desktop-en","Claude Code now controls your Mac desktop","2026-03-28T03:01:59.384091+00:00",{"id":102,"slug":103,"title":104,"created_at":105},"254405b6-7833-4800-8e13-f5196deefbe6","cloudflare-100x-faster-ai-agent-sandbox-en","Cloudflare’s 100x Faster AI Agent Sandbox","2026-03-28T03:09:44.356437+00:00",{"id":107,"slug":108,"title":109,"created_at":110},"04f29b7f-9b91-4306-89a7-97d725e6e1ba","openai-backs-isara-agent-swarm-bet-en","OpenAI backs Isara’s agent-swarm bet","2026-03-28T03:15:27.849766+00:00",{"id":112,"slug":113,"title":114,"created_at":115},"3b0bf479-e4ae-4703-9666-721a7e0cdb91","openai-plan-automated-ai-researcher-en","OpenAI’s plan for an automated AI researcher","2026-03-28T03:17:42.312819+00:00",{"id":117,"slug":118,"title":119,"created_at":120},"fe91bce0-b85d-4efa-a207-24ae9939c29f","harness-engineering-ai-agent-reliability-2026","Harness Engineering: From Bridle to Operating System, The Missing Link in AI Agent Reliability","2026-03-31T06:36:55.648751+00:00",{"id":122,"slug":123,"title":124,"created_at":125},"7a09007d-820f-43b3-8607-8ad1bfcb94c8","mcp-explained-from-prompts-to-production-en","MCP Explained: From Prompts to Production","2026-04-01T09:24:40.089177+00:00",{"id":127,"slug":128,"title":129,"created_at":130},"116d5ee9-a4f1-4b5a-aac5-5d035dd22bbe","amazon-bedrock-agents-multi-agent-workflows-en","Amazon Bedrock Agents Gets Multi-Agent Workflows","2026-04-01T09:30:30.197685+00:00"]