What to know about the AI models that are jolting Washington

Researchers who have tested Anthropic’s Mythos and OpenAI’s GPT-5.5 say their hacking capabilities are a “game-changer.”
format_indent_decrease
AI Article