Claude Opus 5 findet Sicherheitslücken ähnlich gut wie Mythos 5, sagt Anthropic. Beim Ausnutzen der Lücken sehe es aber anders aus.

After the government’s cybersecurity concerns, the company gave this model more cyber safeguards than the previous Opus model.

Anthropic's latest Claude upgrade targets developers and enterprises with stronger coding, better reasoning efficiency, prompt-cache-friendly tool changes, and near-Fable…