---
title: "Previewing the Model Hardware Standard"
description: "Anthropic and HHMI Janelia Research Campus are opening a research preview of the Model Hardware Standard (MHS), a shared specification for AI agents to operate programmable laboratory and manufacturing devices safely. MHS replaces bespoke integrations with standardized drivers built around read and write primitives, device discovery, natural-language tags, generated reference files, and control through MCP, a command-line interface, or code files. The model-agnostic standard lets agents coordinate instruments, monitor operating data, adjust parameters in real time, and turn exploratory work into deterministic scripts; early projects covered protein assays, qPCR, microscopy, robotics, dose-response experiments, and quantum-laser stabilization. Reported results include serial dilution experiments running about three times faster and a QuEra laser-lock controller recovering 99.3% of the time, while the preview will inform safety evaluations and guidance before an open-source release."
---

# Previewing the Model Hardware Standard

[Anthropic](https://yomu.fyi/company/anthropic) · Aug 29, 2026

**Type:** Announcement

## Summary

Anthropic and HHMI Janelia Research Campus are opening a research preview of the Model Hardware Standard (MHS), a shared specification for AI agents to operate programmable laboratory and manufacturing devices safely. MHS replaces bespoke integrations with standardized drivers built around read and write primitives, device discovery, natural-language tags, generated reference files, and control through MCP, a command-line interface, or code files. The model-agnostic standard lets agents coordinate instruments, monitor operating data, adjust parameters in real time, and turn exploratory work into deterministic scripts; early projects covered protein assays, qPCR, microscopy, robotics, dose-response experiments, and quantum-laser stabilization. Reported results include serial dilution experiments running about three times faster and a QuEra laser-lock controller recovering 99.3% of the time, while the preview will inform safety evaluations and guidance before an open-source release.

## Context

Labs and manufacturing facilities often require weeks or months of bespoke integration because devices use separate programming interfaces, lack standardized communication, and have no common method for safe AI-agent control. The preview is intended to support collaboration on safety evaluations and deployment practices before MHS becomes open source.

## Approach / What changed

MHS introduces standardized drivers using read and write primitives, device discovery, natural-language hardware tags, and generated reference files describing capabilities and safety limits. Agents can control connected devices through MCP, a command-line interface, or code files, enabling orchestration, monitoring, real-time parameter changes, and deterministic execution of learned procedures.

## Takeaways

- MHS drivers expose device operations through simple read and write primitives, make hardware discoverable across networks, and provide agents with generated reference files containing characteristics and enforced safety limits.
- Carnegie Mellon researchers used MHS to run serial dilution dose-response experiments about three times faster, coordinating a liquid handler, plate reader, robotic arm, and monitoring cameras across three computers.
- MHS currently requires a programmable hardware interface, and physical reasoning limitations mean expert oversight remains necessary; the research preview will be used to develop additional safety evaluations and protections.

**Tags:** [AI Agents](https://yomu.fyi/topic/ai-agents), [Automation](https://yomu.fyi/topic/automation), [MCP](https://yomu.fyi/topic/mcp), [Robotics](https://yomu.fyi/topic/robotics)

- Source: [Anthropic](https://www.anthropic.com/news/model-hardware-standard-research-preview)
- Source URL: https://www.anthropic.com/news/model-hardware-standard-research-preview
- Ingested by Yomu: 2026-09-09T11:35:15.131Z

[Read original post](https://www.anthropic.com/news/model-hardware-standard-research-preview)