Anthropic expands its most powerful AI models for security teams

Anthropic is expanding a programme that allows verified cybersecurity professionals to test its most powerful artificial intelligence models with fewer safeguards, after its Project Glasswing initiative helped uncover more than 100,000 software vulnerabilities this year.

Its partners in Glasswing, an initiative aimed at securing the world’s most critical software, found at least 129,000 verified vulnerabilities between April and July. Anthropic’s open-source scanning found a further 5,500 between April and October.

More than 33,000 have so far been assessed as critical or severe. Anthropic said the figures are likely to be underestimated and expects the true impact to be at least five times higher, as the data come from a survey of a limited number of partners.

The renewed Cyber Verification Programme, or CVP, announced on Tuesday combines two programmes Anthropic has run over the past six months.

The first, Glasswing, gave organisations access, for critical software, to Claude Mythos, Anthropic’s family of models most capable in cyber terms. The second, the original CVP, gave verified security teams reduced safeguards for the Claude Opus and Sonnet models.

The company’s introduction of Claude Mythos Preview in April raised fears that artificial intelligence could hack software before it was secure.

The new programme has three tiers, each with its own verification requirements and security controls. All three include access to the Claude Opus 5.5, Sonnet 5.5 and Mythos 5.1 models, and to future models.

The Defence Tier covers work such as incident response and malware analysis. Security teams, critical infrastructure operators, open-source maintainers and researchers with a record of reported vulnerabilities may apply.

The Red Team Tier adds authorised penetration testing and red teaming, but only organisations may apply.

The Specialised Tier has fewer restrictions and is reserved for a small group of organisations authorised to test safety-critical systems, such as power grids, flight systems and interbank transfer infrastructure.

Anthropic reviews every member together with the US government, and existing Glasswing members will move to this tier.