Previewing the Model Hardware Standard
Anthropic, in collaboration with HHMI Janelia Research Campus, is opening a research preview of the Model Hardware Standard (MHS), a shared specification designed to let AI agents safely operate physical devices. The preview is initially available to a select group of scientific research labs and advanced manufacturers. MHS is positioned as a solution to a long-standing pain point: lab and factory hardware typically takes weeks or months to set up and integrate, because most devices do not communicate with each other and require specialists to build bespoke integrations.
MHS addresses this with a standardized driver—software that translates between a computer's operating system and a hardware device. The driver uses a simple set of primitives, such as "read" (e.g., get temperature) and "write" (e.g., set temperature), that any device with a programmable interface can understand. It also makes each device discoverable in a standard format, so devices and agents can find each other across networks without needing a bespoke translator. The driver is model-agnostic, works with any agent harness via standard protocols like the Model Context Protocol, and can give an AI agent information about machine characteristics not apparent from code alone (for example, the weight of a robot arm), helping the agent operate hardware it has never seen before.
According to Anthropic, MHS reduces integration work from weeks or months to hours or minutes, and enables AI agents to operate multiple instruments—such as microscopes, liquid handlers, and robotic arms—in parallel, performing tasks ranging from routine drug discovery experiments to laser calibration on a quantum computer. That allows researchers to orchestrate autonomous, round-the-clock experiments, with agents reasoning through each step, updating parameters in real time, and in some cases recovering from hardware errors without human intervention. The early version is being shared with partners across science, robotics, electronics, and manufacturing to collaboratively build safety evaluations and best practices, ahead of making the standard open source.
MHS is designed to work with any device that has a programmable interface and is not tied to any particular AI model, meaning it could become a common layer for agentic control of physical equipment across industries. Interested parties can apply for access to the research preview.