Anthropic limits access to Mythos, its vulnerability-hunting AI model, to 200 partners

Anthropic PBC says its Mythos AI model is so effective at finding vulnerabilities in software and computer systems that the company has decided against a public release. As of early June 2026, only 200 partner organizations had access to Mythos.

What Mythos can do

The model specializes in finding vulnerabilities in software and computer systems, work the cybersecurity industry calls security research or penetration testing. According to Anthropic, Mythos does this so well that the company judged a broad public release too risky.

• Developer: Anthropic PBC
• Mythos's specialty: finding vulnerabilities in software and computer systems
• As of early June 2026, 200 partner organizations have access to the model
• The company does not plan a public release for a wide audience

Why the company restricted access

Vulnerability-hunting tools are dual-use by nature: the same findings that help defenders patch holes in their systems become a ready-made attack plan in the wrong hands. Anthropic points to exactly this risk to justify the restrictions. According to the company, a tool this powerful could make it easier to steal data or disrupt critical infrastructure if it reaches the wrong people.

Instead of an open launch, Anthropic chose a controlled-access approach: only vetted partner organizations are testing Mythos, not the broader community of developers or researchers.

What it means

Mythos illustrates a dilemma AI labs increasingly face when building cybersecurity tools: the better a model finds vulnerabilities for defenders, the more dangerous it becomes in attackers' hands. Anthropic took the cautious route, opting for limited distribution rather than risking a tool of this caliber amplifying cyberattacks on critical infrastructure.

FAQ

How many organizations have access to Mythos?

As of early June 2026, Anthropic has granted access to Mythos to only 200 partner organizations; the model is not available to the general public.

Why didn't Anthropic release Mythos publicly?

The company worries that in the wrong hands, such an effective vulnerability-hunting tool could help attackers steal data or disrupt critical infrastructure.