Google Joins Anthropic’s $104M Project Glasswing to Uncover Zero-Day Vulnerabilities
Google is among roughly 45 firms joining Anthropic’s Project Glasswing cybersecurity initiative, which leverages the new Claude Mythos Preview model. Anthropic is committing $100 million in usage credits and $4 million to open-source groups after Claude Mythos identified thousands of zero-day vulnerabilities, including bugs over two decades old.
1. Partnership Launch
Anthropic has formed Project Glasswing with Google, Amazon, Apple, Microsoft, Nvidia and about 40 other critical software firms to address the rise of AI-powered cyberattacks. Participating companies will share access to Claude Mythos Preview under controlled conditions to strengthen collective defenses.
2. Claude Mythos Preview Capabilities
Claude Mythos Preview, a frontier AI model limited to selected partners, has already detected thousands of zero-day vulnerabilities across major operating systems and web browsers. Highlights include identifying a 27-year-old bug and a 16-year-old flaw previously unknown to developers.
3. Financial Commitments
Anthropic is providing $100 million in usage credits and $4 million in direct donations to support open-source cybersecurity organizations. These resources aim to accelerate vulnerability research and enhance global software security.
4. Implications for Google
By integrating Claude Mythos Preview into its security protocols, Google will proactively detect and patch critical software flaws in products like Chrome and Android. This partnership enhances Google’s defenses against evolving AI-driven cyber threats.