Anthropic rolls out 3-tier access for cyber models
Anthropic is folding its Cyber Verification Program and Project Glasswing into one offering, with three tiers of access to its most capable models for security teams that can prove who they are.
Anthropic has consolidated its Cyber Verification Program and Project Glasswing into a unified offering with three tiers of access to its advanced models for verified security teams.
What happened
Anthropic has integrated its Cyber Verification Program and Project Glasswing into a single offering, introducing three tiers of access to its most capable models for security teams that can verify their identity. The new structure includes Defense Access, Red Team Access, and First-party Access, each with different verification requirements and security controls.
Why it matters
The three-tier access model allows Anthropic to provide secure and controlled access to advanced AI models for cybersecurity professionals, ensuring that defensive uses are supported while maintaining safety measures. This approach addresses the challenge of balancing AI capabilities with security risks, particularly for teams needing to analyze malware, trace exploits, and create detection rules.
- Anthropic's three-tier access model balances AI capabilities with security by verifying user identity and adjusting access levels.
- The initiative supports defensive cybersecurity uses while maintaining safety measures against potential misuse.
- The program's success depends on efficient verification processes and enforcement of real-time security blocks.
The bigger picture
The new access model provides a formal pathway for security vendors and open-source maintainers to use advanced AI models for defense, potentially shifting who can access frontier systems. It also pressures competitors to develop similar programs and influences enterprise purchasing decisions regarding model access and security.
Anthropic's consolidation of its cybersecurity programs into a three-tier access model aims to provide secure and controlled access to advanced AI models for verified security teams, addressing the need for defensive AI use without compromising safety.
Technical details
- Three-tier access model: Defense Access, Red Team Access, and First-party Access.
- Verification requirements and security controls vary by tier.
- Real-time blocking of actions that could cause physical harm or mass disruption.
- Enterprise Frontier Safeguards allow regulated customers to store Claude logs in their cloud environments.
FAQ
What is the purpose of Anthropic's three-tier access model?
The purpose is to provide secure and controlled access to advanced AI models for verified security teams, balancing AI capabilities with safety measures to support defensive cybersecurity uses.
What are the three tiers of access in the new program?
The three tiers are Defense Access, Red Team Access, and First-party Access, each with different verification requirements and security controls.
How does the new model address security concerns?
By verifying user identity and adjusting access levels, the model ensures that defensive uses are supported while maintaining real-time blocks on harmful actions like ransomware deployment.
This summary and analysis were generated with AI from the original article and may contain errors. Read the original for the full story.
Read the full story
Originally published by Dev.to AI on Oct 9, 2026.
Comments
Sign in to join the discussion.