跳到主内容
@wquguru
精选88MarkTechPost(RSS)产品发布/更新

微软开源Agensh:去中心化千级Agent协作框架

Microsoft Open-Sources Agensh: A Multi-Agent Harness for 1,024 Agents

原文
发到 X
推荐理由

Agent编排是当下热点,Agensh提供去中心化千级扩展的一手实践与完整代码,对构建大规模Coding Agent团队极具参考价值。

Microsoft Research has open-sourced Agensh, a decentralized harness for large teams of coding agents. The Agensh multi-agent harness removes the central orchestrator that most agent frameworks depend on. Every worker picks its own sub-task, shares findings and merges code into a common Git repo. On ProgramBench, the team scaled the setup from 1 agent to 1,024 agents.

微软研究院开源了 Agensh,这是一个面向大型编码智能体团队的分布式框架。Agensh 多智能体框架移除了大多数智能体框架所依赖的中心化编排器。每个工作节点自主选择子任务,共享发现结果,并将代码合并到公共 Git 仓库中。在 ProgramBench 上,团队将设置从 1 个智能体扩展到 1,024 个智能体。

TL;DR

TL;DR(太长不看)

  • A Harness: Tested from 1 to 1,024 agents. Each worker gets a 272,000-token input and 128,000-token output limit.
  • Runs on: Linux with Docker Compose. Single node by default (8 workers). The 1,024-agent run used 16 nodes, 64 agents each. Workers have a 64 GiB memory ceiling.
  • Performance: more agents raised scores on all 5 hardest ProgramBench tasks, and reached them sooner.
  • Best: 79.59% pandoc pass rate with 32 agents on GPT-6.1-sol. GPT-5.6-sol hit 55.06% with 1,024 agents.
  • Worst: the 5-task mean reached only 28.78% at 128 agents. Going from 1 to 8 agents added 1.37 points.
  • 一个框架:经过 1 到 1,024 个智能体的测试。每个工作节点获得 272,000 token 的输入和 128,000 token 的输出限制。
  • 运行环境:基于 Docker Compose 的 Linux。默认单节点(8 个工作节点)。1,024 智能体运行使用了 16 个节点,每节点 64 个智能体。工作节点的内存上限为 64 GiB。
  • 性能表现:增加智能体数量提高了所有 5 个最难的 ProgramBench 任务的得分,并更早达到这些分数。
  • 最佳表现:在 GPT-6.1-sol 上使用 32 个智能体时,pandoc 通过率达到 79.59%。GPT-5.6-sol 在使用 1,024 个智能体时达到 55.06%。
  • 最差表现:5 个任务的平均值在 128 个智能体时仅为 28.78%。从 1 个增加到 8 个智能体增加了 1.37 分。

What is Agensh?

什么是 Agensh?

Agensh is a multi-agent harness from Microsoft Research that coordinates many LLM agents without a lead agent.

Agensh 是微软研究院推出的多智能体框架,它协调多个 LLM 智能体,无需主智能体。

The target problem is the orchestrator ceiling. In orchestrator-worker designs, a lead agent plans, assigns and merges all work. The research paper cites Claude Code agent teams, Codex sub-agents, Copilot fleet and Kimi Agent Swarm as examples. Agensh lets the workers do that coordination themselves.

目标问题是“编排器瓶颈”。在编排器-工作节点设计中,主智能体负责规划、分配和合并所有工作。研究论文引用了 Claude Code 智能体团队、Codex 子智能体、Copilot 舰队和 Kimi Agent Swarm 作为例子。Agensh 让工作节点自行完成这种协调。

How does the Agensh cooperation loop work?

Agensh 的合作循环如何运作?

Each worker runs the same 5-step loop, concurrently and asynchronously:

每个工作节点并发且异步地运行相同的 5 步循环:

  • Gather context: read the goal, peer progress, messages and findings.
  • Claim a sub-task: post a CLAIM, then settle overlaps by direct message.
  • Take action: work locally with tools and post useful findings early.
  • Verify results: check the work against the sub-task’s acceptance criteria.
  • Merge progress: merge into the shared repo, post an update, then repeat.
  • 收集上下文:读取目标、同伴进度、消息和发现结果。
  • 认领子任务:发布 CLAIM,然后通过私信解决重叠问题。
  • 采取行动:使用工具在本地工作,并尽早发布有用的发现。
  • 验证结果:根据子任务的验收标准检查工作成果。
  • 合并进度:合并到共享仓库,发布更新,然后重复。

Three services connect the workers. Gitea hosts the shared workspace, with branches, PRs and merge conflicts. Mattermost carries a team channel plus direct messages for urgent collisions. A shared context board, adapted from DeLM, stores typed notes: OBSERVED, FACT, FAIL, CLAIM and PATCH_SUMMARY.

三项服务连接工作节点。Gitea 托管共享工作区,包含分支、PR 和合并冲突。Mattermost 承载团队频道以及用于紧急冲突的直接消息。共享上下文板改编自 DeLM,存储类型化的笔记:OBSERVED(已观察)、FACT(事实)、FAIL(失败)、CLAIM(认领)和 PATCH_SUMMARY(补丁摘要)。

The loop lives in the prompt, not in runtime code. Every worker gets the same prompt except its ID. Agensh sits on top of a single-agent harness. The experiments used GitHub Copilot CLI. The research paper states thin adapters can connect other harnesses, such as Claude Code.

该循环位于提示词中,而非运行时代码中。每个工作节点获得相同的提示词,仅 ID 不同。Agensh 建立在单智能体框架之上。实验使用了 GitHub Copilot CLI。研究论文指出,轻量级适配器可以连接其他框架,例如 Claude Code。

How well does Agensh scale on ProgramBench?

Agensh 在 ProgramBench 上的扩展效果如何?

Microsoft team tested the 5 hardest ProgramBench tasks: FFmpeg, gromacs, pandoc, PHP-src and ctags. Agents rebuild each program from its observed behavior. They work offline, within 6 hours. Reference repos range from 2,767 files (pandoc) to 26,266 files (PHP-src).

微软团队测试了 ProgramBench 中最难的 5 个任务:FFmpeg、gromacs、pandoc、PHP-src 和 ctags。智能体根据观察到的行为重建每个程序。它们在离线环境下运行,耗时不超过 6 小时。参考仓库的文件数量从 2,767 个(pandoc)到 26,266 个(PHP-src)不等。

With GPT-5.6-sol at high reasoning effort, the mean final test-pass rate moved as follows:

在高推理努力设置下,使用 GPT-5.6-sol,最终测试通过率的变化如下:

  • 1 agent: 19.31%
  • 8 agents: 20.68%
  • 32 agents: 26.52%
  • 128 agents: 28.78%
  • 1 个智能体:19.31%
  • 8 个智能体:20.68%
  • 32 个智能体:26.52%
  • 128 个智能体:28.78%

That is a 9.47-point gain, about 49% relative. On pandoc alone, scores were 33.89% (1), 41.75% (8), 43.11% (32), 50.94% (128) and 55.06% (1,024).

这带来了 9.47 个百分点的提升,相对增幅约为 49%。仅在 pandoc 上,得分分别为 33.89%(1 个)、41.75%(8 个)、43.11%(32 个)、50.94%(128 个)和 55.06%(1,024 个)。

Speed improved too: on pandoc, 128 agents passed 30% at the 30-minute checkpoint. 32 and 8 agents needed 60 and 90 minutes.

速度也有所提升:在 pandoc 上,128 个智能体在 30 分钟检查点时通过率达到 30%。而 32 个和 8 个智能体分别需要 60 分钟和 90 分钟。

The README adds a newer model. With GPT-6.1-sol, pandoc rose from 56.19% (1 agent) to 79.59% (32 agents). That is a 41.64% relative gain. These GPT-6.1-sol figures are not in the v1 paper.

README 中增加了一个更新的模型。使用 GPT-6.1-sol,pandoc 的得分从 56.19%(1 个智能体)提升至 79.59%(32 个智能体)。相对增益为 41.64%。这些 GPT-6.1-sol 的数据未包含在 v1 论文中。

What behaviors emerged as the team grew?

随着团队规模的扩大,出现了哪些行为模式?

The research team traced worker trajectories at each scale:

研究团队追踪了不同规模下的工作轨迹:

  • 8 agents: peers agreed on interfaces and resolved overlapping claims themselves.
  • 32 agents: peers reviewed shared work; one approval was withdrawn after a counterexample.
  • 128 agents: workers reused trusted reviewers and standardized an integration protocol.
  • 1,024 agents: several workers acted as integrators, and peers recovered failed work.
  • 8 个智能体:同行就接口达成一致,并自行解决重叠的主张。
  • 32 个智能体:同行审查共享的工作;在一个反例出现后,一项批准被撤回。
  • 128 个智能体:工作者复用受信任的审查者,并标准化了集成协议。
  • 1,024 个智能体:部分工作者充当集成者,同行则恢复失败的工作。

None of these roles was hard-coded.

这些角色均未硬编码。

How does Agensh compare with other multi-agent harnesses?

Agensh 与其他多智能体框架相比如何?

FeatureAgenshClaude Code agent teamsKimi Agent SwarmCopilot CLI /fleet
CoordinationNo orchestrator; peers self-claim workFixed lead session coordinates teammatesOrchestrator decomposes tasksCopilot assigns parts to subagents
Agents shown1,024 testedNo hard limit; 3 to 5 recommendedUp to 100 sub-agentsNot disclosed
Shared stateGitea repo + shared context boardShared task list with file lockingNot disclosedPlan from plan mode
MessagingMattermost channel + direct messagesMailbox messages between teammatesNot disclosedNot disclosed
ModelsModel agnostic; tested GPT-5.6-sol, GPT-6.1-solClaude modelsKimi K2.5Not disclosed
Speed claim30% on pandoc at 30 min (128 agents)Not disclosed4.5x faster than sequential (self-reported)Not disclosed
AvailabilityOpen source, MITIn Claude Code; experimental, off by defaultTop-tier subscribers, research previewIn Copilot CLI
特性AgenshClaude Code 智能体团队Kimi Agent SwarmCopilot CLI /fleet
协调方式无编排器;同行自主认领工作固定的主导会话协调队友编排器分解任务Copilot 将部分分配给子智能体
展示的智能体数量测试了 1,024 个无硬性限制;建议 3 到 5 个最多 100 个子智能体未披露
共享状态Gitea 仓库 + 共享上下文看板带有文件锁定的共享任务列表未披露来自计划模式的计划
消息传递Mattermost 频道 + 直接消息队友之间的邮箱消息未披露未披露
模型模型无关;测试了 GPT-5.6-sol、GPT-6.1-solClaude 模型Kimi K2.5未披露
速度声明128 个智能体在 30 分钟内对 pandoc 达到 30% 通过率未披露比顺序执行快 4.5 倍(自报)未披露
可用性开源,MIT 协议位于 Claude Code 中;实验性,默认关闭顶级订阅用户,研究预览版位于 Copilot CLI 中

How do you run Agensh?

如何运行 Agensh?

The release includes a single-node Docker Compose team and a single-agent baseline. Both use native Copilot ACP sessions. You need Linux with Docker Engine, Compose v2, systemd, cgroup v2 and Python 3.11+. Defaults are 8 workers, 6 hours per task and gpt-5.6-sol at high effort. You also need ProgramBench (v1.2.5 was validated) and a model API token. Full steps sit in the reproduction guide.

该版本包含一个单节点 Docker Compose 团队和一个单智能体基线。两者均使用原生的 Copilot ACP 会话。你需要配备 Docker Engine、Compose v2、systemd、cgroup v2 和 Python 3.11+ 的 Linux 系统。默认设置为 8 个工作者、每项任务 6 小时以及高努力程度的 gpt-5.6-sol。你还需要 ProgramBench(已验证 v1.2.5)和一个模型 API 令牌。完整步骤请参阅复现指南。

Key Takeaways

关键要点

  • Agensh coordinates up to 1,024 agents with no central orchestrator.
  • Workers self-claim tasks via Gitea, Mattermost and a shared findings board.
  • 128 agents lifted the 5-task ProgramBench mean from 19.31% to 28.78%.
  • GPT-6.1-sol with 32 agents hit 79.59% on pandoc.
  • Code is MIT licensed, but token cost is not disclosed.
  • Agensh 协调多达 1,024 个智能体,无需中央编排器。
  • 工作者通过 Gitea、Mattermost 和共享的发现板自主认领任务。
  • 128 个智能体将 ProgramBench 上 5 项任务的平均值从 19.31% 提升至 28.78%。
  • GPT-6.1-sol 配合 32 个智能体在 pandoc 上达到了 79.59%。
  • 代码采用 MIT 许可证,但未披露令牌成本。

Check out the Paper, GitHub Repo and Project Page. Also, feel free to follow us on Twitter and don’t forget to join our 150k+ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well.

请查看论文、GitHub 仓库和项目页面。此外,欢迎在 Twitter 上关注我们,别忘了加入我们拥有 15 万+成员的 ML SubReddit,并订阅我们的新闻通讯。等等!你在 Telegram 上吗?现在你也可以在 Telegram 上加入我们。

更进一步:量化金融体系

看懂新闻只是起点——沿量化金融路径,把它变成能交付的工程能力

进入量化体系 →

相似阅读

关联信息,但可能不是同一事件