AWS has announced the availability of Claude Fable 5.1 on Amazon Bedrock and Claude Platform on AWS. Anthropic positions the model for demanding coding, scientific research, and enterprise workflows.
Because Fable 5.1 is designated a Covered Model, its availability includes additional data-retention, safety-review, and access policies. Anthropic also announced Enterprise Frontier Safeguards, designed to let eligible customers deploy advanced models while retaining control of their cloud infrastructure.
Capabilities
Anthropic reports that Fable 5.1 improves on Fable 5 in competition mathematics, graduate-level engineering and science questions, and long, multi-step reasoning tasks. The company says the model makes better decisions on ambiguous work and produces fewer confident errors.
- Agentic coding: Handles codebase-wide features, code reviews, and performance work across multi-hour sessions, while being more likely to acknowledge when it is stuck.
- Autonomous operation: Plans multi-hour jobs across applications, uses tools, recovers from failures, and provides progress updates.
- End-to-end knowledge work: Performs research, creates spreadsheets, writes memos or presentations, and checks calculations, with applications in finance, accounting, and healthcare.
- Scientific research: Supports work from literature reviews and hypotheses through models, experiments, and formal verification.
- Usability: Provides clearer writing, closer instruction following, and more updates during lengthy tasks.
Data retention
As a Covered Model, Claude Fable 5.1 may be subject to retention for up to 30 days and human review by Amazon personnel. On Amazon Bedrock, customers configure retention through the data retention API at the account or project level.
Invoking Fable 5.1 requires the aws_review mode. AWS retains prompts and outputs within the AWS boundary for human safety review, and the model does not require sharing data with its provider.
Enterprise Frontier Safeguards
Enterprise Frontier Safeguards, developed by AWS and Anthropic, will allow eligible customers to use Claude Fable 5 and Claude Fable 5.1 with zero data retention on Amazon Bedrock and Claude Platform on AWS. Internal use is available through December 31, 2026.
Later in 2026, additional safeguards are planned to keep prompts and outputs in a customer’s AWS account under its own encryption keys, access policies, and audit logging. Safety monitoring will use automated review without human review.
Getting started
In the Amazon Bedrock console, users can open Test > Playground, select Fable 5.1, and submit a prompt. Developers can call it through the Anthropic Messages API using bedrock-runtime and the Anthropic SDK, or use the Invoke and Converse APIs through the AWS CLI and AWS SDK.
Requirements include an active AWS account with Amazon Bedrock access, a configured AWS CLI installation, Python 3.10 or later, Boto3 installed with pip install boto3, and IAM permissions for bedrock:InvokeModel and bedrock:InvokeModelWithResponseStream.
A Boto3 request uses the model ID global.anthropic.claude-fable-5-1 through the bedrock-runtime client. Since Fable 5.1 is a reasoning model, its response can contain a thinking block before the text block; applications should select the content block whose type is text rather than relying on a fixed array position.
Availability
Claude Fable 5.1 is available on Amazon Bedrock through US Geo CRIS and Global CRIS inference profiles. In AWS GovCloud (US), it is available through both the bedrock-runtime and bedrock-mantle endpoints. It is also offered through Claude Platform on AWS in North America.