What Is Anthropic’s Mythos AI and Why Was It Blocked?

Anthropic has restricted access to its new vulnerability-scanning AI, Mythos, limiting the tool to 200 selected partners rather than a public release. The company stated that the model’s high proficiency in identifying software security flaws poses a significant risk if used maliciously. By gatekeeping the technology, Anthropic aims to prevent potential attackers from exploiting the same automated reconnaissance capabilities that security researchers use to harden systems.