Microsoft Developing 'Project Perception,' a Multi-Model AI Security Tool to Rival Anthropic's Mythos

Microsoft is building an AI-powered cybersecurity product, reportedly called Project Perception, that would route vulnerability-discovery and remediation tasks across its own, OpenAI's and Anthropic's models to offer enterprises a lower-cost alternative to Anthropic's high-end Mythos system. The effort, described in reporting from The Information, could debut as early as this month, though Microsoft has not made an official announcement.