Anthropic expands its Mythos model usage to approximately 150 new organizations Foresight News reported that Anthropic announced the expansion of its "Project Glasswing" project. The project aims to use the AI model Claude Mythos Preview to scan for security vulnerabilities in the code repository. Since the first batch of about 50 partners joined in early April this year, these institutions have cumulatively discovered over 10000 high-risk or serious security vulnerabilities. This expansion will expand the scope of cooperation to about 150 new organizations, covering more than 15 countries, involving key infrastructure industries such as electricity, water, healthcare, communication, and hardware that were previously underrepresented. \Anthropic stated that newly joined partners must meet their security requirements in order to obtain access, and the common feature of all partners is that once their code repository is successfully attacked, it may affect more than 100 million people and have a significant impact on global and national security. At the same time, Anthropic also announced the opening of its internal development tools to trusted security teams and the launch of a new product called "Claude Security" based on the cutting-edge public model Claude Opus 4.8, which is used to scan code repositories and recommend patches. Anthropic expects multiple AI companies to launch models with the same level of network security capabilities in the next 6 to 12 months, and has stated that it will continue to expand the project in order to establish long-term advantages for network defense.