Anthropic Grants METR Employee-Level Access for Independent Safety Evaluation

ChainCatcher report: On October 15, Anthropic announced it has begun granting employee-like access to a team at the nonprofit METR to evaluate its models and safety practices. The company stated that last month it committed to enhancing independent third-party evaluations and signed the White House Superintelligence Agreement, which includes a pledge to collaborate with independent external auditors or evaluators. Over the past year, Anthropic has worked with METR to review several risk reports, and METR will publish its risk assessment findings. Anthropic has provided six METR employees with company hardware, office space, and internal tools and system access roughly equivalent to those available to its own risk assessment team. Except for a few clearly defined categories—such as customer data, attorney-client privileged information, and model architecture details—employees are encouraged to share information as they would with colleagues working on alignment risks. METR may publish analyses of risk levels, incidents, or practices; Anthropic cannot veto publication but may request the removal of narrowly predefined categories of information; METR may publicly disclose that redactions occurred and, without revealing the redacted content, state whether it considers the redactions reasonable. This arrangement is limited to a small team operating under an enforceable agreement, with other METR staff kept isolated from this team’s work. Anthropic is also collaborating with SecureBio and Accenture on evaluations and is in early discussions with other organizations, including Hugging Face. SecureBio is currently reviewing the chemical and biological weapons section of the August 2026 risk report, having previously completed and published a pilot review of the prior risk report.