{"version":"1.0","generated_at":"2026-11-06T09:50:55.427410","id":3815,"slug":"no-ai-agent-scores-above-12-percent-on-semianalysis-new-real-world-work-benchmark","title":"SemiAnalysis 新真实工作基准：没有一个 AI agent 得分超过 12%","summary":"分析机构 SemiAnalysis 发布 GDPx 基准，把 AI agent 放进模拟职场环境，让它们连续数小时操作专业软件、完成 206 项任务。最高分只有 12%（Claude Opus 5.5），Gemini 4 Carbon 以 10.9% 紧随其后但单项成本约为前者的三分之一，GPT-6 Astra 为 9.6%。批评者质疑安全过滤、任务保密与各家自有 agent 软件造成的不平等条件，让结果更像噪声。","abstract":"SemiAnalysis 新真实工作基准：没有一个 AI agent 得分超过 12% 分析机构 SemiAnalysis 发布 GDPx 基准，把 AI agent 放进模拟职场环境，让它们连续数小时操作专业软件、完成 206 项任务。最高分只有 12%（Claude Opus 5.5），Gemini 4 Carbon 以 10.9% 紧随其后但单项成本约为前者的三分之一，GPT-6 Astra 为 9.6%。批评者质疑安全过滤、任务保密与各家自有 agent 软件造成的不平等条件，让结果更像噪声。 GDPx 让 agent 在模拟职场连续跑 8 小时、完成 206 项任务 榜首 Claude Opus 5.5 仅 12%，无模型超过该线 Gemini 4 Carbon 10.9%，单项成本约为榜首三分之一 200 次网络攻击测试中 98% 至少违反一条行为规则 任务不公开可防刷榜，但外部难验证难度与评分质量 分析公司 SemiAnalysis 发布了 GDPx，这是一个在真实工作环境和专业软件中、持续数小时测试 AI agent 的基准，涵盖 206 项任务。分数很低。Claude Opus 5.5 以仅 12% 领先，其后是谷歌价格显著更低的 Gemini 4 Carbon（10.9%）和 GPT-6 Astra（9.6%）。批评者指出安全过滤器带来的失真、测试任务的保密性，以及各家自有 agent 软件造成的不平等条件。 SemiAnalysis 的新 GDPx 基准把 AI agent 放进模拟的知识工作中。没有模型得分超过 12%。一些批评者质疑这项测试到底有多大…","access_level":"member","access_label":"免费阅读","access_mode":"full","is_preview":false,"canonical_url":"https://opc.beizhux.com/content/3815/no-ai-agent-scores-above-12-percent-on-semianalysis-new-real-world-work-benchmark","html_url":"https://opc.beizhux.com/content/3815/no-ai-agent-scores-above-12-percent-on-semianalysis-new-real-world-work-benchmark","json_url":"https://opc.beizhux.com/content/3815/no-ai-agent-scores-above-12-percent-on-semianalysis-new-real-world-work-benchmark.json","published_at":"2026-11-06T00:46:24","updated_at":"2026-11-06T09:35:45","category":{"slug":"ai-business","name":"AI小生意项目库"},"source":{"site":"the-decoder.com","author":"Manuel Uth","url":"https://the-decoder.com/no-ai-agent-scores-above-12-percent-on-semianalysis-new-real-world-work-benchmark/"},"tags":["AI Agent","GDPx","SemiAnalysis","企业落地","基准测试","成本效率","模型评测"],"topics":[{"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":"agent-workflow","name":"Agent工作流","url":"https://opc.beizhux.com/topics/agent-workflow","reason":"这篇内容命中「Agent」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「模型、Claude」等主题信号。"}],"keywords":["AI小生意项目库","AI副业","Agent工作流","AI工具","工具","自动化","模型","Cursor","Claude","Agent","智能体","工作流"],"questions":[{"question":"SemiAnalysis 新真实工作基准：没有一个 AI agent 得分超过 12%主要讲什么？","answer":"分析机构 SemiAnalysis 发布 GDPx 基准，把 AI agent 放进模拟职场环境，让它们连续数小时操作专业软件、完成 206 项任务。最高分只有 12%（Claude Opus 5.5），Gemini 4 Carbon 以 10.9% 紧随其后但单项成本约为前者的三分之一，GPT-6 Astra 为 9.6%。批评者质疑安全过滤、任务保密与…"},{"question":"这篇文章最值得关注的要点是什么？","answer":"分析机构 SemiAnalysis 发布 GDPx 基准，把 AI agent 放进模拟职场环境，让它们连续数小时操作专业软件、完成 206 项任务。最高分只有 12%（Claude Opus 5.5），Gemini 4 Carbon 以…；GDPx 让 agent 在模拟职场连续跑 8 小时、完成 206 项任务；榜首 Claude Opus 5.5 仅 12%，无模型超过该线；Gemini 4 Carbon 10.9%，单项成本约为榜首三分之一"},{"question":"这篇文章和哪些AI专题相关？","answer":"它适合放在AI副业、Agent工作流、AI工具专题里阅读。 关联原因：这篇内容命中「项目、小生意、变现」等主题信号。；这篇内容命中「Agent」等主题信号。；这篇内容命中「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":3809,"title":"Copilot 用量指标 API 新增字段，标记遥测缺失与版本不达标","url":"https://opc.beizhux.com/content/3809/usage-metrics-api-flags-telemetry-and-version-gaps","summary":"GitHub Copilot 用量指标 API 新增逐用户字段，可直接看出哪些用户没有上报客户端遥测、哪些用户的 IDE 或扩展版本低于用量指标的最低要求，并在企业与组织报告中给出每日受影响人数，便于定位和补齐数据缺口。","access_level":"member"},{"id":3811,"title":"深入解析 GitHub 全新的 Git 存储架构","url":"https://opc.beizhux.com/content/3811/inside-github-s-new-git-storage-architecture","summary":"GitHub 推出新的 Git 存储架构，目标是支撑数千个智能体同时向同一个仓库推送代码，官方同步说明了其工作方式。","access_level":"member"},{"id":3812,"title":"Google Gemini Omni 1.2 Flash 将 AI 视频片段延长至 15 秒，价格降低 30%","url":"https://opc.beizhux.com/content/3812/google-s-gemini-omni-1-2-flash-stretches-ai-video-clips-to-15-seconds-and-cuts-pricing-by-30-percent","summary":"Google DeepMind 宣布 Gemini Omni 1.2 Flash 视频生成模型全面可用。该模型可生成最长 15 秒的片段，支持对话式编辑，价格比前代低 30%。","access_level":"member"}],"usage_policy":{"summarizable":true,"preferred_url":"https://opc.beizhux.com/content/3815/no-ai-agent-scores-above-12-percent-on-semianalysis-new-real-world-work-benchmark","private_fields_excluded":true}}