Anthropic's Mythos AI: Public Release & Cybersecurity Impact

Anthropic plans to eventually release a public version of Mythos, its AI model with advanced cybersecurity capabilities. However, the company admits it first needs to develop adequate safeguards, which do not yet exist. Mythos was initially revealed in April. During testing, it generated working exploits 72.4 percent of the time. This compares to just above zero percent for Claude Opus 4.6. Since then, the model has scanned over a thousand open-source projects, identifying more than 23,000 flaws. Over 6,000 of these were rated high or critical severity. One vulnerability found in the wolfSSL cryptography library, used by billions of devices, has since been patched. Open-source maintainers have asked Anthropic to slow its disclosure rate because the flood of bug reports exceeds their capacity. Anthropic suggests more AI tools could help with this capacity crunch. This development could significantly change the landscape of cybersecurity for everyone.