Anthropic opens Model Hardware Standard research preview
Anthropic opened an early version of its specification for connecting AI agents with programmable laboratory and manufacturing equipment.
Quick answer
What did Anthropic announce about the Model Hardware Standard?
Anthropic opened a research preview of the Model Hardware Standard, a shared specification for AI agents to operate programmable physical devices. The company says MHS standardizes communication with equipment including microscopes, liquid handlers and robotic arms, and is initially available to selected research labs and advanced manufacturers.
Key takeaways
- Anthropic opened the Model Hardware Standard research preview to an initial group of scientific research labs and advanced manufacturers.
- The company says MHS provides a shared specification through which AI agents can operate multiple programmable physical devices.
- MHS uses standardized drivers and supports control through the Model Context Protocol, a command-line interface and APIs.
- Anthropic says the preview will support work with partners on safety evaluations and best practices before the standard becomes open source.
- Genentech tested MHS and Claude in a proof of concept involving a liquid handler, robotic arm and microplate reader.
Anthropic has opened a research preview of the Model Hardware Standard, or MHS, to an initial group of scientific research laboratories and advanced manufacturers. The company describes MHS as a shared specification that allows AI agents to operate physical devices safely.
Anthropic says the standard can connect agents with multiple laboratory and manufacturing instruments, including microscopes, liquid handlers and robotic arms. Agents can operate those devices in parallel for work ranging from routine drug-discovery experiments to calibrating a laser on a quantum computer. Development began through a collaboration between Anthropic and HHMI Janelia Research Campus.
A common interface for programmable devices
MHS introduces a standardized software driver that translates between a computer operating system and a hardware device. The driver uses basic commands such as “read” and “write,” while describing each connected device in a standard format so devices and agents can discover one another and communicate across networks.
Users can add natural-language tags containing information about equipment, including characteristics that may not be evident from its code. Anthropic says those tags can record details such as the weight of a robotic arm. The driver then generates a reference file describing what the device can measure, which settings can be changed and which safety limits will be enforced.
Agents can control connected equipment through the Model Context Protocol, a command-line interface or code files using APIs. According to Anthropic, these mechanisms allow an agent to coordinate several devices with one line of code. The agent can sequence operations, monitor results and change parameters as conditions develop. For longer or faster operations, it can assemble driver commands into code files that devices execute without requiring the agent to reason through every step.
Anthropic says MHS is model-agnostic, works with any device that has a programmable interface and can be used by any agent harness through standard protocols. The company is sharing the early version with partners in science, robotics, electronics and manufacturing to develop safety evaluations and operating practices before making the standard open source.
Tests with Claude and laboratory equipment
In Anthropic’s testing, Claude adjusted a laser, assessed the beam’s movement through a camera and repeated the process. The company says Claude then converted what it learned into a deterministic script that aligned the laser through a single command.
Genentech also implemented MHS in a proof of concept for automating a BCA protein assay, a procedure used to measure total protein concentration. The setup connected a liquid handler, a robotic arm and a microplate reader, with Claude coordinating the protocol.
According to the material published by Anthropic, Claude tested liquid-transfer flow rates for water and a viscous bovine serum albumin solution, then analyzed readings from the microplate reader against an expert-performed transfer. Genentech reported that Claude selected about 140 microliters per second for water and 10 microliters per second for the protein solution. It also encountered tip-pickup and fluid-detection errors and recovered without intervention, although the experiments showed limitations in handling physical, chemical and biological constraints.
Source: Anthropic’s “Previewing the Model Hardware Standard” announcement, published Aug. 27, 2026.
Frequently asked questions
- What is the Model Hardware Standard?
- Anthropic describes MHS as a shared specification for AI agents to operate physical devices safely. It works with devices that have programmable interfaces.
- Who can access the preview?
- Anthropic opened the research preview to a first group of scientific research labs and advanced manufacturers and provided an application process for access.
- Which devices can MHS control?
- Anthropic lists microscopes, liquid handlers and robotic arms as examples. The company says MHS works with any device that has a programmable interface.
- Does MHS require Claude?
- Anthropic says MHS is model-agnostic and can be accessed by any agent harness using standard protocols such as the Model Context Protocol.