Anthropic Launches Restricted-Function Model Fable 5, Disabling Cybersecurity-Related Capabilities

TechFlow news: On June 10, Bloomberg reported that Anthropic broadly launched its new model Fable 5 on Tuesday. Fable 5 incorporates safety guardrails that prohibit responses to requests related to cybersecurity and biology; such queries are instead routed to the Claude chatbot, which handles them using the Opus 4.8 model.

Meanwhile, Anthropic also released a new version, Mythos 5, which retains more capabilities but is available exclusively to specific institutions covered under Project Glasswing. Anthropic previously stated that Mythos—under user guidance—can identify and exploit vulnerabilities in major operating systems and mainstream web browsers.