Anthropic’s Claude Code and Security Tool Nears Public Release

Key Takeaways

• Anthropic is preparing for the commercial release of Claude Mythos 1, its most powerful AI model, integrating it into Claude Code and Claude Security.
• Mythos 1, initially restricted to Project Glasswing partners, is now poised for a wider, staged commercial rollout, specifically targeting enterprise customers.
• The AI model has already identified over 10,000 high- or critical-severity vulnerabilities in open-source software during its restricted phase.
• Claude Security is being developed into a comprehensive dashboard for vulnerability scanning, patch suggestions, and historical analysis.

Anthropic’s Claude Mythos 1 Nears Public Debut in Enterprise Security and Code Tools

Anthropic is reportedly gearing up for the commercial launch of Claude Mythos, its highly anticipated and most advanced AI model. Recent indicators suggest that a version, designated Mythos 1 (claude-mythos-1-preview), will be integrated directly into the company’s enterprise offerings: Claude Code and a re-engineered Claude Security dashboard.

The existence of Claude Mythos first came to light in late March 2026, following an accidental data exposure from Anthropic’s content management system. This leak revealed nearly 3,000 internal documents, including a draft blog post that described the model as “by far the most powerful AI model we’ve ever developed.”

Days after the leak, Anthropic officially acknowledged the model, characterizing its capabilities as a “step change,” particularly in the domains of cybersecurity, coding, and academic reasoning.

Instead of an immediate general release, Anthropic initiated Project Glasswing on April 7, 2026. This restricted defensive cybersecurity program granted exclusive access to over 40 partner organizations, including industry giants such as AWS, Apple, Google, Microsoft, NVIDIA, CrowdStrike, and JPMorgan Chase.

During its deployment within Project Glasswing, the model was tasked with scanning critical infrastructure. Within weeks, it successfully identified more than 10,000 high- or critical-severity vulnerabilities across a wide array of commonly used open-source software projects.

Mythos 1: Indications of a Phased Commercial Release

Anthropic’s strategy for Mythos is now signaling a shift towards broader availability. An updated Project Glasswing report indicates that the initiative has expanded its protective scope to include open-source projects. Crucially, the report states that “Mythos-class models could reach the public once the right safeguards are in place,” a significant departure from earlier statements that implied the model would remain indefinitely restricted.

More concrete evidence emerged from source code strings referencing claude-mythos-1-preview. These strings were accompanied by a new access description: “Access to the Claude Mythos model in Claude Code and Claude Security.”

Further confirmation came from Testingcatalogue, which shared a screenshot of the Mythos 1 model label within the Claude user interface before its removal. This brief appearance validates that the product name is actively being prepared for release, rather than being a hypothetical concept.

This development firmly positions Mythos 1 for integration into two key Anthropic products: Claude Code, the company’s agentic developer environment, and Claude Security, a vulnerability scanning and patch-suggestion tool currently in public beta for enterprise clients.

Claude Security initially launched in February 2026 as a limited research preview, powered by Claude Opus 4.7. Its core function involved scanning codebases and proposing targeted patches for human review. The current beta remains exclusive to Claude Enterprise customers, with future access for Team and Max subscribers having been announced.

Significant architectural work is underway to introduce a new dashboard interface for Claude Security. This dashboard will display discovered vulnerabilities, alongside seven-day and thirty-day historical charts, and provide deeper triage results.

Evidence within the Claude navigation bar points to the development of a standalone Security tab, positioned alongside Claude Code and Claude Design. This move suggests Anthropic is packaging the scanner into a more accessible, consumer-style dashboard, moving beyond direct enterprise sales conversations.

The underlying architecture is designed to support periodic automated background scans, on-demand manual scans, and the generation of pull requests with proposed patches, ready for human review before merging into codebases.

Anthropic’s coordinated vulnerability disclosure dashboard, last updated on May 22, 2026, underscores the scale of these efforts. It reflects 1,596 vulnerabilities disclosed across 281 open-source projects, with 97 patches implemented to date.

Concurrently, rumors have circulated about Claude Opus 4.8 undergoing internal evaluation with select Anthropic partners. This would align with the established release cadence, following Opus 4.7’s launch on April 16.

Anthropic’s own rollout of Opus 4.7 incorporated cybersecurity safeguards designed to automatically detect and block high-risk offensive use cases. These measures were explicitly framed as a preparatory step towards a broader release of Mythos-class models.

The emerging strategy depicts a deliberate, layered commercialization approach: Mythos 1 will initially be a gated capability within enterprise products. This will be followed by a measured expansion, contingent on the establishment of robust safety evaluations and infrastructure safeguards to manage the model’s formidable and potentially dangerous cyber capabilities.

Disclaimer: HackersRadar reports on cybersecurity threats and incidents for informational and awareness purposes only. We do not engage in hacking activities, data exfiltration, or the hosting or distribution of stolen or leaked information. All content is based on publicly available sources.