Table of Contents
Anthropic said it is expanding access to Claude Mythos 5, its most capable AI model, through new Claude Security scanning capabilities, partnerships with cybersecurity tool makers and a $35 million fund for open-source security work, marking a significant broadening of Claude Mythos 5 cyber defenders access since the model’s restricted debut.
What Claude Security Brings to This Claude Mythos 5 Cyber Defenders Expansion
Claude Security, now in public beta for Claude Enterprise customers, can run codebase scans on Claude Mythos 5 to find vulnerabilities and suggest patches, according to an Aug. 21 blog post. Enterprise admins can enable the feature in the admin console, and every suggested patch requires human review before a team implements it.
Why Human Review Remains a Required Step
Keeping human review mandatory before any suggested patch is implemented ensures that even as Claude Mythos 5’s vulnerability-finding capabilities expand to more organizations, the actual decision to modify production code remains firmly in the hands of human security teams rather than being automated end to end.
How Partner Tools Extend Claude Mythos 5 Cyber Defenders Access
Anthropic is also building Claude Mythos 5 into products from cybersecurity technology and services partners that security teams already use for alert triage, threat detection and remediation. Users of those partner tools won’t interact with Mythos directly; they will receive only the specific defensive output, such as a list of suggested patches, the company said.
Why Indirect Access Matters for This Expansion Strategy
By routing Mythos’s capabilities through existing partner tools rather than granting direct model access, Anthropic can extend the model’s defensive value to a broader base of security teams already embedded in familiar workflows, while maintaining tighter control over how the underlying model itself is accessed and used.
The Defender Advantage Fund Within This Claude Mythos 5 Cyber Defenders Initiative
The new Defender Advantage Fund will provide $35 million in Claude credits to organizations that patch vulnerabilities in widely used open-source projects, automate scanning and patching, or pursue more resilient security approaches. Anthropic plans to name initial grant recipients in the coming weeks.
Why Open-Source Security Represents a Strategic Priority
Directing this fund specifically toward widely used open-source projects reflects a recognition that vulnerabilities in shared, foundational software components can have outsized downstream effects across countless organizations that depend on the same open-source code, making concentrated investment in this area a potentially high-leverage use of AI-driven security resources.
Expanding the Cyber Verification Program to Include Mythos-Class Access
Anthropic also said it will expand its Cyber Verification Program, which gives vetted security teams reduced safeguards on Claude Opus and Sonnet, to include Mythos-class access in the coming weeks. This extension gives already-vetted security teams a pathway toward more capable model access without requiring an entirely separate vetting process.
How This Builds on Project Glasswing
The moves build on Project Glasswing, the restricted-access program Anthropic launched in April to let a limited set of organizations, including some in healthcare, use Mythos to find and fix software flaws before those capabilities become widely available to attackers. Epic Systems has publicly confirmed its own participation in Glasswing, using the model to surface vulnerabilities across its several-hundred-million-line codebase.
What This Claude Mythos 5 Cyber Defenders Expansion Means Going Forward
With Claude Security now in public beta and the Cyber Verification Program set to expand to Mythos-class access in the coming weeks, this rollout suggests Anthropic is moving methodically from Project Glasswing’s initial restricted cohort toward progressively broader availability for vetted defensive use cases. Given healthcare organizations’ documented participation in Project Glasswing and health IT vendors like Epic already applying these capabilities to their own codebases, hospital and health system security teams may find increasing opportunities to access similar AI-driven vulnerability detection as these programs expand.
What to Watch Going Forward
As Anthropic names initial Defender Advantage Fund grant recipients in the coming weeks, industry observers will likely watch which organizations and open-source projects receive support, and whether the expanded Cyber Verification Program meaningfully increases the number of healthcare and other critical-infrastructure organizations gaining Mythos-class access. Given the parallel dynamic between AI capabilities accelerating both defenders’ and attackers’ work, this Claude Mythos 5 cyber defenders expansion may shape how quickly healthcare organizations and other security teams can close vulnerability windows before adversaries gain access to comparably capable tools.
For more healthcare industry updates, insights and news, visit DistilINFO. Click here to subscribe to stay informed.
