Claude Fable 5: Anthropic's Safe AI for Cybersecurity

Anthropic has launched Claude Fable 5, a new Mythos-class AI model, now generally available. This model includes specific safeguards to limit its use in high-risk areas like cybersecurity. What's interesting is that this is the first time an AI of this capability class is considered safe for widespread public and developer access. Fable 5 performs well in software engineering and knowledge work, but Anthropic prioritized safety. Here's the thing: in sensitive domains such as cybersecurity and biology, the model automatically switches to the less capable Claude Opus 4.8. This prevents potential misuse. Early data shows that 95% of sessions use Fable 5's full capabilities without this fallback. Anthropic conducted extensive internal and external testing, including a bug bounty program, and found no universal ways to bypass these safety measures. The company also announced that partners in Project Glasswing, including cybersecurity firms, are being upgraded to Claude Mythos 5. Both Fable 5 and Mythos 5 are priced at $10 per million input tokens and $50 per million output tokens. This development shows a focus on balancing powerful AI with strong safety protocols for critical applications.