Select Language:
Anthropic has announced the launch of a new initiative called Project Glasswing, aimed at bolstering the security of critical global software infrastructure using a yet-to-be-released model named Claude Mythos Preview. The company is currently limiting access to this program to vetted organizations, which include major industry players such as Amazon, Apple, Broadcom, Cisco, Google, Microsoft, Nvidia, and the Linux Foundation.
To support these efforts, Anthropic is providing participating organizations with a total of $100 million worth of model usage credits. This funding allows these entities to utilize the Claude Mythos Preview model for conducting security assessments on their foundational systems.
Initial testing results have been promising, with reports indicating that the model has automatically identified thousands of high-risk vulnerabilities within mainstream operating systems and browsers. Notably, among these are a 27-year-old vulnerability in OpenBSD and a 16-year-old flaw hidden within FFmpeg, which had previously escaped detection even after millions of automated scans.
In addition, Anthropic has committed to releasing a public report within 90 days detailing the vulnerabilities found and offering best practices for remediation. These recommendations include automated patching and supply chain security enhancements. The company also noted ongoing discussions with U.S. government agencies regarding the model’s capabilities in both offensive and defensive cybersecurity operations, emphasizing its potential role in strengthening digital defenses across sectors.
ChatGPT
Perplexity
Gemini AI
Grok AI
