Mythos-class

Anthropic released Claude Fable 5, its most powerful publicly available AI model, but it avoids answering basic biology questions by redirecting to the older Claude Opus 4.8 model. The restriction is by design, as Fable is part of the Mythos-class models deemed too dangerous for cybersecurity tasks, though it retains the capability to answer such questions.

Anthropic's Claude Fable 5 model, part of the 'Mythos-class' series, includes broad safety safeguards that can block or revert to Opus 4.8 when queries about cybersecurity, biology, or chemistry are flagged. The model's restrictions were implemented to prevent potential misuse in sensitive areas, even if they mistakenly block benign requests.

Anthropic released Claude Fable 5, a 'Mythos-class' AI model with safeguards restricting cybersecurity, biology, chemistry, and distillation queries to its Opus 4.8 model. The model exceeds previous Anthropic models in capabilities but is limited in access, with a 'trusted access program' planned for a version without full safeguards.