As artificial intelligence reshapes the cybersecurity landscape, major technology companies are collaborating to counter emerging AI driven threats. Anthropic has launched Project Glasswing, a large scale initiative backed by 100 million dollars in model usage credits to identify and fix vulnerabilities across global software infrastructure. The Project Glasswing AI security initiative brings together industry leaders including Apple, Microsoft, Amazon Web Services, Broadcom, Cisco, Google, Nvidia, and Linux Foundation. The coalition aims to deploy advanced AI models to scan, test, and strengthen both proprietary and open source systems against increasingly sophisticated cyber threats.
At the core of the initiative is Claude Mythos Preview, a next generation AI model developed by Anthropic. The model is designed for advanced reasoning and code analysis, with the capability to autonomously identify software vulnerabilities at a scale that exceeds traditional methods. According to Anthropic, early testing revealed that the model was able to uncover thousands of high severity vulnerabilities across major operating systems and web browsers, including long standing flaws that had remained undetected for years.
Among the findings were a 27 year old vulnerability in OpenBSD and a 16 year old flaw in FFmpeg, both of which had previously passed numerous automated security checks. These results highlight the growing role of AI in vulnerability discovery and underscore the need for equally advanced defensive capabilities.
To support the initiative, Anthropic is providing 100 million dollars in model usage credits to participating organizations. The company is also contributing an additional 4 million dollars to open source security efforts, including funding for Alpha Omega, OpenSSF, and the Apache Software Foundation. These investments are intended to strengthen the broader software ecosystem and improve resilience across shared infrastructure.
The Project Glasswing AI security initiative will not make Claude Mythos Preview publicly available. Instead, access will be limited to vetted participants through platforms such as the Claude API, Amazon Bedrock, Google Cloud Vertex AI, and Microsoft Foundry. Pricing for usage has been set at 25 dollars per million input tokens and 125 dollars per million output tokens.
Anthropic has indicated that the project will run for several months, with a public report expected within 90 days. This report will detail vulnerabilities discovered during the initiative and provide actionable recommendations for improving security practices, including automated patching and stronger supply chain protections. The company has also engaged in discussions with U.S. government officials regarding both the defensive and potential offensive implications of such advanced AI capabilities.
The Project Glasswing AI security initiative represents a significant step toward coordinated, AI driven cybersecurity defense. As AI continues to evolve, collaborative efforts like this will be essential to identifying risks early, securing critical infrastructure, and ensuring that defensive technologies keep pace with rapidly advancing threats.
Recommended Cyber Technology News :
- Pluralsight Launches SecureReady for Cyber Skills Gap
- AI Supply Chain Attack Targets GitHub Repositories
- Nothreat Partners With AZAL To Strengthen Aviation Cybersecurity
To participate in our interviews, please write to our CyberTech Media Room at info@intentamplify.com
