SAN FRANCISCO, Calif. — Anthropic has launched a major defensive cybersecurity initiative called “Project Glasswing,” forming a coalition of leading technology companies to proactively secure global software infrastructure. The effort leverages the company’s unreleased “Claude Mythos Preview,” an advanced AI model capable of autonomously discovering and chaining together high-severity software vulnerabilities. The details of the project were extensively reported by The Washington Post, CSO Online, Silicon Republic, Security Affairs, CTV News, and MacRumors. Each of the bullet points immediately below have been confirmed by at least four of the six respected sources we curated on this story.
- Anthropic launched the “Project Glasswing” initiative as a collaborative effort to enhance global cybersecurity defenses alongside major industry partners.
- The project is powered by an unreleased frontier artificial intelligence model known internally as “Claude Mythos Preview.”
- The advanced AI model has demonstrated the ability to autonomously identify, analyze, and exploit thousands of high-severity software vulnerabilities.
- Anthropic has deliberately withheld the model from public release due to significant concerns over its potential misuse by malicious threat actors.
- To support defensive operations, the company is dedicating up to $100 million in usage credits to allow partner organizations to scan and secure their own systems.
Additional Details Reported
Project Glasswing represents a strategic pivot in AI deployment, attempting to position frontier models as defensive shields before they can be weaponized. The coalition includes major enterprises such as Amazon Web Services, Apple, Broadcom, Cisco, and Microsoft, highlighting the industry-wide urgency to identify flaws that have previously evaded detection. In addition to the usage credits provided to corporate partners, Anthropic announced $4 million in direct donations to open-source security organizations to shore up the foundational software ecosystems that underpin much of the modern internet. The cautious approach to releasing Claude Mythos Preview underscores the growing recognition of AI’s dual-use potential in the cybersecurity landscape.
Image Attribution ▾
Attribution: Royalty-free photo fallback utilized (Hedra image generation failed due to API key permission restrictions)
How we report: We select the day’s most important stories, confirm facts across multiple reputable sources, and avoid anonymous sourcing. Our goal is clear, balanced coverage you can trust—because transparency and verification matter for informed readers.