Anthropic's Claude Mythos outperforms OpenAI's GPT-5.5 in UK cyber simulations
Technology
The UK AI Security Institute reported that Anthropic's latest AI, Claude Mythos, outperformed both its older version and OpenAI's GPT-5.5 in UK cyber simulations.
The new Mythos Preview handled tricky problems that stumped earlier models, showing how fast AI is leveling up.
Mythos succeeded under 2.5 million token limit
Mythos pulled this off even with a 2.5-million-token testing limit, hinting it could be even better if those restrictions were lifted.
These quick upgrades aren't just about beating competitors; they're pushing what AI can do, especially for future cybersecurity challenges.