Original sourceMarkTechPost
Summary
Microsoft Research has open-sourced Agensh, a decentralized multi-agent collaboration framework that can coordinate up to 1,024 coding agents without a central orchestrator. In ProgramBench tests, Agensh scaled from 1 to 1,024 agents, with each worker node handling 272k input and 128k output tokens…
Key points
- Teams needing large-scale parallel coding agents can draw on its decentralized coordination, shared repository and memory budget design.
- It demonstrates a viable architecture for coordinating over a thousand coding agents without a central scheduler, with reproducible benchmark data.
- Teams looking to scale agentic coding throughput can assess self-deploying on Docker Compose and Git workflows, but must first estimate node and memory costs.
Editorial note
This page is Code & Chain's editorial summary of public sources. It may be prepared with AI assistance and published through an automated workflow. Refer to the original sources; this content is not investment, legal, or tax advice.