Anthropic deploys Mythos 5 in security tools, launches $35M open-source fund

Anthropic announced on Friday that it is committing $35 million in Claude credits to a new fund for securing open-source software, while simultaneously deploying its most capable cybersecurity model, Claude Mythos 5, into enterprise security scanning tools for the first time.

Mythos 5 Goes Live in Claude Security

The Defender Advantage Fund, dubbed 0xDAF, will distribute grants to organizations patching vulnerabilities in widely used open-source projects, automating scanning and patching workflows, and pursuing approaches that make software resistant to entire classes of attack. Anthropic said it will start with a small number of larger pilot grants and name initial recipients in the coming weeks.claude

Alongside the fund, Claude Enterprise customers can now run vulnerability scans powered by Claude Mythos 5 directly from claude.ai/security. The feature, billed as standard token usage with no add-on fee, returns findings tagged with Common Weakness Enumeration categories, confidence and severity ratings, and suggested fixes. Every patch must be reviewed and approved by a human before implementation.unite+1

Building on Project Glasswing

The moves extend a deliberate rollout that began in April 2026 with Project Glasswing, a coalition including AWS Amazon.com, Inc. ↘ 0.57%, Microsoft ↗ 0.43%, Google Alphabet Inc. ↗ 1.05%, CrowdStrike ↗ 0.85%, and the Linux Foundation. That initiative committed up to $100 million in usage credits and $4 million in direct donations — $2.5 million to Alpha-Omega and the Open Source Security Foundation through the Linux Foundation, and $1.5 million to the Apache Software Foundation.anthropic+1

Anthropic reported in April that Mythos Preview had found thousands of high-severity vulnerabilities, including a 27-year-old remotely triggerable flaw in OpenBSD and a 16-year-old bug in FFmpeg. Reuters reported in June that approximately 200 organizations had received access under the Glasswing program.Reuters+1

Expanding Access Without Expanding Risk

Anthropic is also working with cybersecurity partners — firms such as Wiz, Palo Alto Networks ↗ 2.38%, and CrowdStrike that have already built products on Claude Opus — to integrate Mythos 5 into their existing tools. End users of those products interact with purpose-built interfaces rather than prompting the model directly, receiving only specific artifacts like patch suggestions.unite+1

The company's Cyber Verification Program, which currently gives vetted defenders reduced safeguards on Opus and Sonnet models, will expand in coming weeks to include Mythos-class access for defensive capabilities like vulnerability triaging and validation. Anthropic framed the approach as widening access to defensive results while maintaining guardrails: "The riskiest behavior occurs when a user has direct access to a model," the company wrote. "But if users can only receive specific outputs, such as a patch for a vulnerability or a security alert, that risk is much lower."claude