Anthropic Brings Claude Mythos 5 to Claude Security

Anthropic has announced a major update, bringing its advanced Claude Mythos 5 model to Claude Security for enterprise customers in a public beta release. There is no separate model add-on; Mythos 5 scans are billed as standard token usage under existing organizational plans. Administrators can enable the feature directly in the admin console, allowing users to initiate scans at claude.ai/security.

Along with this announcement, Anthropic is collaborating with cybersecurity vendors to embed Mythos 5 inside partner tooling, with interest forms currently open. Users can connect their GitHub repositories, enabling Claude Mythos 5 to scan code by tracing data flows across files and reading Git history rather than relying on traditional rule-based pattern matching.

Stock photo for illustration only, not from the actual event

Each discovered finding undergoes an adversarial verification step where the model challenges its own results before presentation, serving as the core mechanism for reducing false positives. Findings are returned with CWE categories, confidence and severity ratings, and suggested fixes targeting high-severity issues such as memory corruption, injection flaws, authentication bypasses, and cross-file logic errors.

"Anthropic has stated Mythos 5 has the strongest cybersecurity capabilities of any model in the world, and that the same skills that find vulnerabilities can write exploits."

Anthropic

Results can be seamlessly integrated via webhooks to Slack or Jira, or exported as CSV or Markdown files. Scans can also be scoped to specific directories and scheduled in advance. Patching operates separately through Claude Code on the web using existing organization models, and every patch strictly requires human review and approval.