Anthropic Fits Mythos 5 into Claude Security: Enterprise-Ready, But Model Not Available

Dynamic Beating AI Newsflash: Anthropic has upgraded Claude Security's underlying scanning model to Claude Mythos 5 and opened public testing to all Claude Enterprise customers. Once an enterprise connects to the GitHub repository, Mythos 5 will scan for code vulnerabilities and provide CWE classification, confidence level, severity rating, and repair suggestions.

Claude Security was actually launched some time ago. When it opened for public testing at the end of April, it was still using Claude Opus 4.7. The biggest change this time is integrating the previously strictly restricted Mythos 5 into the existing security product. Mythos 5 not only identifies vulnerabilities but can also convert them into actionable exploits, hence it has always been open only to audited institutions.

Enterprises still cannot directly access Mythos 5. The model only runs in the scanning backend, with users receiving vulnerability reports and repair suggestions. Subsequently, modifications are made in Claude Code, still using the enterprise's existing model, and patches must be manually approved. Scans are billed with existing packages according to the Token system and do not require a separate purchase of Mythos 5.

Anthropic is also planning to integrate Mythos 5 into partner security products and launch the Defender Advantage Fund with a $35 million Claude allocation to help open-source projects identify and patch vulnerabilities.