Anthropic launches Claude Fable 5 with safeguards, limits Mythos 5 access

Anthropic released a public version of Mythos while barring high-risk uses such as cybersecurity, and said Claude Fable 5 is its strongest broadly deployable model so far.

Summary

Anthropic has launched a public version of its Mythos model while barring use in high-risk areas such as cybersecurity, and said Claude Fable 5 is its strongest broadly deployable model so far. The company highlighted software engineering and data analysis performance and said both Mythos preview and Claude Mythos 5 are priced at $10 per million input tokens and $50 per million output tokens. The rollout extends Anthropic's tiered deployment strategy, pairing broader access with safeguards and tighter limits around more sensitive capabilities.

Terms & Concepts
  • input tokens: Units of text or data sent into an AI model for processing and billing.
  • output tokens: Units of text generated by an AI model that are commonly used for pricing.
  • high-risk areas: Sensitive use cases that a company restricts because they may pose elevated safety or security concerns.