Anthropic may soon release Claude Mythos 1 security testing AI
Technology
Anthropic might be gearing up to release its new AI, Claude Mythos 1, soon.
Clues hidden in its code and security tools (and even on Google Cloud and AWS) suggest the launch is close.
This model is all about finding software vulnerabilities and could be available through Claude Code.
Claude Mythos 1 flagged 10,000+ vulnerabilities
Claude Mythos 1 has already flagged over 10,000 high- or critical-severity vulnerabilities since April, making it Anthropic's most advanced model yet.
The public release was paused because of worries about misuse: Anthropic says no one has cracked the code on perfect safety yet.
If launched safely, this AI could seriously boost how companies protect themselves online, but everyone is watching to see how risks are managed.