Anthropic Releases First Public Mythos-Class Model Claude Fable 5

Tradingkey - Anthropic announced on Tuesday the launch of Claude Fable 5, the company's first Mythos-class AI model to be widely available to enterprise customers and paid users.

Two months ago, the same-tier Mythos model was released only to a limited number of users due to safety concerns. This broad release is primarily attributed to new protection mechanisms for high-risk domains, which can block responses to dangerous queries in areas such as cybersecurity and biology.

Claude Fable 5 shows significant performance improvements in software engineering and knowledge work tasks, with some benchmark scores more than 10% higher than those of Claude Opus 4.8 released last month.

For high-risk questions, the Fable 5 model features built-in rigorous safety protections that block queries in high-risk areas like cybersecurity and biology, automatically switching to the Opus 4.8 model to generate a safe response when such issues are encountered.

Anthropic also launched Claude Mythos 5, which shares the same underlying architecture but has some safety restrictions removed, continuing the previous model of being limited to specific scenarios.

Mythos 5 removes certain safety safeguards; this version will be made available through "Project Glasswing" to institutions qualified to use models with cybersecurity capabilities.

Anthropic reportedly added 150 institutions to the Mythos access list last week, bringing the total number of authorized institutions to approximately 200.