NewsBB-0179Curator pick

Anthropic previews the Model Hardware Standard for agents running lab equipment

Anthropic opened a research preview of the Model Hardware Standard, a shared driver spec for agents that operate lab and manufacturing devices. Each device describes its settings and enforced safety limits, and agents reach it over MCP, a CLI or code. In a QuEra case study, an agent-improved laser relock script worked in 99.3% of 700 trials.

Open original source anthropic.com · Anthropic

Claimed
Why this evidence label?

Announcement text with partner-reported results; no logs or videos in the post.

How we check entries
Preview image for Anthropic previews the Model Hardware Standard for agents running lab equipment
The source's own preview image, from anthropic.com.

Briefing

MHS gives each device a standard driver with read and write primitives, network discovery and a description of what it measures, what can be adjusted and which safety limits apply. Anthropic developed it with HHMI Janelia, and it does not depend on any one model.

The partner case studies give numbers. QuEra's relock script, refined by the agent overnight, recovered the laser lock 695 times in 700 trials; the script a human team built had worked about 58% of the time. Carnegie Mellon induced six faults, among them a missing plate and an active emergency stop, and the system blocked all six before any device moved. Janelia replaced seven vendor programs on one rig.

The results are the partners' own. Access is through a waitlist, and Anthropic says it will open-source the standard after the preview.

“MHS works with any device that has a programmable interface.”

Anthropic, in the source on anthropic.com

Tags

Agent
Interface

Report a problem with this entry