Microsoft said it plans to integrate advanced AI models — including Claude Mythos Preview from Anthropic — into its secure coding framework, as it strengthens its cybersecurity capabilities.
The models will be embedded within Microsoft’s Security Development Lifecycle (SDL), enabling developers to detect vulnerabilities earlier and accelerate the process of fixing them, according to the company.
Mythos, unveiled earlier this month, has already identified thousands of significant vulnerabilities across operating systems, web browsers and other software. Its advanced coding capabilities allow it not only to detect weaknesses but also to simulate how they could be exploited — raising both opportunities and concerns within the industry.
Anthropic is initially rolling out Claude Mythos Preview to a select group of companies under its “Project Glasswing” programme, with participants including Amazon and Apple alongside Microsoft.
Microsoft said its internal testing, using proprietary benchmarks for real-world detection scenarios, showed “substantial improvements” compared to earlier AI models.
The rapid advancement of tools like Mythos is drawing attention from governments and regulators worldwide, including the administration of Donald Trump, as concerns grow over the potential for AI to make sophisticated cyberattacks faster and easier to execute.
