Anthropic deploys Claude Mythos 5 for cyber defense tools

Anthropic is now running its Claude Security scanner on Claude Mythos 5. The tool scans codebases for vulnerabilities, provides severity ratings, assigns CWE classifications, and suggests patches. Claude Security is available in public beta for Enterprise customers, and scans count as normal token usage.

The scanner can suggest fixes, but the article states that humans still have to sign off on every patch. That keeps the model positioned as an assistant for security teams rather than an autonomous patching system. For organizations evaluating the tool, the key workflow question is how suggested fixes are reviewed, tested, and approved before deployment.

Anthropic is also plugging Mythos 5 into partner security products used to protect hospitals, utilities, and banks. End users will not interact with the model directly; they will see outputs such as suggested patches. Several partners already use Claude Opus for security tools and are expected to switch to Mythos 5.

Claude Mythos is described as Anthropic’s most capable model, especially for cyber tasks. The rollout is framed as a way to boost defenders without giving attackers new AI-powered options. Security vendors can sign up for partnership access, according to the article.