Claude 现在能并行调度最多 1000 个 AI 智能体：同一份代码库的 70 个隐藏缺陷，单个智能体最多找出 27 个，多智能体工作流稳定找到 66 个
Anthropic's Claude can now orchestrate up to 1,000 AI agents in parallel through dynamic workflows
精选评分 78
导读
Anthropic 为 Claude 的托管智能体加上动态工作流：一个主智能体可以一次性把任务分派给最多 1000 个子智能体并行执行。文章的实测口径很具体：在一项测试里，单个智能体面对某代码库中隐藏的 70 个缺陷，最多只找出 27 个；而采用多智能体工作流后，稳定找出的数量是 66 个。这组对照是判断并行编排实际价值时少见的量化依据——它同时说明单智能体在长任务上的覆盖能力有天花板，而这一代产品把「拆给很多个」做成了官方能力。
推荐理由
难得给出可对比的单/多智能体量化结果，把「并行编排到底值不值」这一问回答了一半。
主体与标签
Anthropic#产品更新#Agent#评测/基准
正文
Anthropic is adding dynamic workflows to Claude Managed Agents, letting a lead agent distribute tasks across up to 1,000 sub-agents at once. In testing, a single agent found at most 27 of 70 hidden bugs in a codebase, while the multi-agent workflow consistently caught 66.
The article Anthropic's Claude can now orchestrate up to 1,000 AI agents in parallel through dynamic workflows appeared first on The Decoder .