AWS Launches Anthropic's Claude Fable 5.2 and Mythos 5.2 with New Enterprise Safety Framework
AWS announced the availability of Anthropic's latest models, Claude Fable 5.2 and the verification-gated Mythos 5.2, on Amazon Bedrock, introducing Enterprise Frontier Safeguards (EFS) to balance high-capability agent deployment with strict data compliance.
Amazon Web Services (AWS) announced on November 9, 2026, that two new models from Anthropic, Claude Fable 5.2 and Claude Mythos 5.2, are now available on Amazon Bedrock and the Claude Platform on AWS. This release marks an expansion of enterprise generative AI capabilities into coding, scientific computing, and high-risk specialized domains like cybersecurity and life sciences, while introducing new data governance mechanisms to address compliance challenges.
Claude Fable 5.2 is positioned as Anthropic's most capable generally available model for technical work, designed to handle ambiguous multi-step tasks, verify its own output, and run unattended for hours or days. Its use cases include agentic coding across large codebases, long-running coordination of subagents, scientific research, and financial analysis. In contrast, Claude Mythos 5.2 focuses on cybersecurity and life sciences, available only through Anthropic's verification programs for authorized users performing sensitive tasks such as red teaming, penetration testing, and biomolecular design.
A key change in this update involves data retention policies and security frameworks. Anthropic has designated Claude Fable 5.2 as a 'Covered Model,' meaning its use requires data retention for up to 30 days and human safety review by default. On Amazon Bedrock, this review is performed by AWS personnel within the AWS boundary. To balance privacy and security, AWS introduced 'Enterprise Frontier Safeguards' (EFS). EFS combines the privacy benefits of Zero Data Retention (ZDR) with the latest safeguards for detecting misuse. Currently, eligible customers can use Fable 5.2 with ZDR through June 30, 2027. Additionally, AWS previewed a version allowing customers to store data in their own buckets under keys they control, though this feature has not yet launched.