Anthropic released Claude Fable 5.2 and Claude Mythos 5.2. Fable 5.2 is positioned for agentic coding across large codebases, long-running multi-agent work, scientific research, and quantitative financial analysis. Mythos 5.2 is restricted to organizations approved through Anthropic verification programs for cyber security and life sciences research.
On Amazon Bedrock both models are available today through US Geo CRIS and Global CRIS inference profiles. Fable 5.2 is designated a Covered Model requiring up to 30 days retention and human safety review by default. Operators can set aws_review data retention mode to keep review inside the AWS boundary.
Eligible customers can use Fable 5.2 with zero data retention for internal use through at least June 30, 2027, until Enterprise Frontier Safeguards become available. A fallback credit token is issued when a declined request hits the prompt cache so the retry is billed only as a cache read.
Anthropic states Fable 5.2 can run unattended for hours to days, verify its own output, and orchestrate agent teams. The model is accessible via the Anthropic Messages API on bedrock-runtime or the Invoke and Converse APIs, with model ID global.anthropic.claude-fable-5-2.