Anthropic / Wikimedia Commons (Public domain)
Vetted security teams can now get graduated access to Claude Opus 5.5, Sonnet 5.5 and Mythos models with loosened safeguards for defensive work
Anthropic is opening the vault a little wider. On October 6, 2026, the AI company expanded its Cyber Verification Program (CVP) into a three-tier system, according to a report credited to Axios.
The overhaul folds in Project Glasswing, the company’s existing partnership with some of the largest names in tech. The goal is to give security professionals graduated access to Anthropic’s top models, including Claude Opus 5.5, Sonnet 5.5 and Mythos 5.1.
The catch is the verification step. Reduced cybersecurity safeguards apply only to defensive and red-teaming operations, and only for people and organizations that pass vetting.
How the three tiers work #
The new structure splits access into three levels: Defense Access, Red Team Access and Specialized Access. Each tier is tailored to a different level of risk and a different set of operational needs.
Speed is part of the offer too. Applications for the new tiers will be answered within days, which should shorten the wait for qualified security teams.
The Glasswing numbers #
Project Glasswing originally launched in April 2026 with a select group of partners. Founding members include Amazon Web Services, Apple and Microsoft, with Google and Nvidia also among the companies using advanced AI to defend critical infrastructure.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
Between April and July 2026, Glasswing partners identified at least 129,000 verified software vulnerabilities. Of those, over 33,000 were classified as critical or high severity.
Anthropic’s own scanning added another 5,500 vulnerabilities through October 2026.
Claude Mythos 5 has shown superior capability in vulnerability discovery, but access was initially restricted to limit potential misuse.
Money for open-source security #
Anthropic is also putting cash behind the effort. The company allocated $100M in usage credits and $4M in donations to support open-source security work.
This follows the $35M Defender Advantage Fund that Anthropic created in August 2026, aimed at strengthening open-source security initiatives.
The company has framed the broader effort as collaboration with both industry and government. The stated aim is to bolster national cybersecurity against persistent threats.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our