Anthropic, an American AI company, has announced the launch of two new AI models from its advanced “Mythos-class” tier, Claude Fable 5 and Claude Mythos 5.
In their announcements, they share that Claude Fable 5 is the first model of this class to be made generally available for public use. To mitigate the risks associated with such highly capable AI, Anthropic has integrated conservative safety classifiers into Fable 5. Queries related to high-risk domains like cybersecurity, biology, chemistry, and model distillation will be automatically detected and rerouted to the company’s next-most-capable model, Claude Opus 4.8.
Claude Mythos 5 shares the same underlying architecture and capabilities as Fable 5 but will operate with specific domain safeguards removed. Due to its advanced dual-use capabilities, access to Mythos 5 is strictly restricted.
According to their announcement, both models demonstrate an increased capacity for autonomous, long-running tasks. Anthropic plans to use data from this initial deployment to refine its conservative safeguard thresholds and reduce false positives, while preparing a broader trusted access program to expand Mythos 5 access to verified institutional researchers.
