Anthropic launches cheaper, less restrictive Fable and Mythos 5.1 AI models
Anthropic unveiled Fable and Mythos 5.1, its latest AI models that cost less per token and feature fewer false-positive safeguards.
On Tuesday Anthropic released paired AI models named Fable and Mythos 5.1, delivering upgraded capabilities alongside reduced token costs and fewer overly restrictive safety blocks. Fable 5.1 is immediately usable on major cloud platforms and through the Anthropic API, while Mythos 5.1 stays exclusive to registered partners in cybersecurity or life-science fields. The rollout includes a new high-privacy service, Enterprise Frontier Safeguards, slated for June, which lets customers run models on their own infrastructure under a zero-data-retention policy.
Anthropic assured that it has never trained on enterprise data without explicit consent. Benchmark tests show the models set new records on tasks such as Terminal-Bench 4.0 and Humanity’s Last Exam, and the system card rates Mythos as low-risk for automated AI development, noting a slight increase in misaligned behavior compared with Opus 5 but improvements over earlier versions.
Why it matters
The update makes advanced AI more affordable and flexible for businesses while promising stronger data privacy controls.
In this story
