AnthropicAnalysisAnthropic expands Cyber Verification Program with three frontier access tiers
What happened.Anthropic launched an expanded multi-tiered Cyber Verification Program to provide vetted security organizations with specialized access to frontier models under adjusted safety classifiers.
Why it matters.Standard safety classifiers frequently block legitimate enterprise security teams from utilizing frontier models for vulnerability discovery, automated remediation, and adversary simulation. Formalizing vetted access tiers allows frontier providers to support critical national and corporate cyber defense operations without distributing unrestricted offensive exploitation tools to malicious actors.
The technical specifications:
- Three-tier vetting framework: The program establishes Defense Access, Red Team Access, and Specialized Access tiers tailored to distinct operational scopes across threat modeling, vulnerability research, and critical infrastructure defense.
- Frontier model availability: Approved defense organizations receive access to Claude Opus 5.5, Claude Sonnet 5.5, and Claude Mythos 5.1 with reduced blocking on dual-use cybersecurity inquiries.
- Enterprise workspace governance: Approvals operate strictly at the organizational workspace boundary, requiring multi-factor authentication, audited API credentials, and query telemetry to prevent unauthorized downstream redistribution.
STORYLINE PROGRESSION & CONTEXT
This report represents a verified evolution of prior newsroom intelligence: