Anthropic Launches Claude Science AI Workbench and Grant Program

Anthropic developed the Model Hardware Standard, or MHS, with HHMI Janelia Research Campus to address a longstanding obstacle in laboratories and advanced manufacturing: equipment from different vendors often lacks a common interface. The specification is designed to let AI agents coordinate programmable physical devices, potentially enabling continuous automated workflows in drug discovery, microscopy, quantum computing and manufacturing while reducing the need for bespoke integrations.

Anthropic opened a limited MHS research preview on Aug. 27, 2026, for an initial group of scientific laboratories and advanced manufacturers. The system can operate microscopes, liquid handlers and robotic arms in parallel, supporting tasks from dose-response experiments to laser calibration on a quantum computer. Anthropic said MHS could cut hardware integration from weeks or months to hours or minutes. The company plans to develop safety evaluations with preview partners before releasing the standard as open source.

Anthropic has sought in recent years to move Claude beyond a general-purpose chatbot and into professional research settings, launching the Claude Science AI workspace to help researchers review literature, analyze data and test hypotheses. Its collaboration with San Francisco biomedical research organization Gladstone Institutes focuses on using real-world scientific data to assess whether AI can produce reproducible results rather than merely demonstrate concepts.

The online event ran from July 7 to July 13, 2026, with Anthropic, Gladstone Institutes and Cerebral Valley accepting up to 500 participants across research and development tracks. Competitors used real-world data involving T cells, the regulatory genome and protein interactions, together with Claude Science or Claude Code, to compete for a total of $100,000 in Claude API credits and usage allowances.