tech
Anthropic releases first Mythos-level model for general use
Fable 5 includes safeguards to help protect against hacking and bio-terrorism.
TL;DR
- Anthropic releases Fable 5, a Mythos-class AI model for general use.
- Fable 5 includes safeguards to prevent misuse in hacking or sensitive biological queries.
- High-risk queries are automatically rerouted to the less capable Claude Opus 4.8 model.
- The underlying model for Mythos and Fable 5 is the same.
- Anthropic conducted extensive safety evaluations, red-teaming, and adversarial testing.
- Fable 5 is priced double that of Opus models but offers better performance.
- Mythos-class models previously required significant safeguards for broader access.
- Anthropic is developing a trusted-access program for vetted users of Mythos 5.