Anthropic announced on Tuesday the launch of its collaborative initiative, 'Project Glasswing,' granting partners such as Amazon and Microsoft access to its unreleased general-purpose AI model, 'Claude Mythos Preview,' for defensive cybersecurity efforts. Although not specifically trained for cybersecurity, the model has already identified 'thousands' of critical vulnerabilities in software due to its powerful programming and reasoning capabilities. It can autonomously recognize vulnerabilities and develop exploit code. To prevent misuse, Anthropic has adopted a phased release strategy, allowing access only to partners and approximately 40 key software infrastructure organizations. The company has committed to providing up to $100 million in usage credits and donating $4 million to open-source security organizations. Participating organizations are required to share their findings. This release comes amid growing concerns in the cybersecurity industry about AI-driven attacks, with Anthropic engaged in ongoing discussions with U.S. government officials regarding the model's capabilities. The company was previously involved in legal disputes with the Trump administration.