Today, we are announcing the availability of Claude Fable 5.2 and Claude Mythos 5.2 on Amazon Bedrock and Claude Platform on AWS. Claude Fable 5.2 is Anthropic’s most capable generally available model for technical work that raises the bar in coding and the sciences. Claude Mythos 5.2 is available through Anthropic’s verification programs for advanced work in cyber security and life sciences.

In this post, we cover what’s new in both models, why Fable 5.2 matters for your technical workloads, and how to get started building with the model on Amazon Bedrock.

Why Claude Fable 5.2

Fable 5.2 raises the bar on the hardest problems in coding, math, and science. It takes on ambiguous, multi-step work, verifies its own output, and can run unattended for hours to days. It interprets prompts, iterates, catches its own errors, and optimizes as it goes.

In Anthropic’s alignment testing, Mythos 5.2 hallucinates less, completes tasks more reliably, surfaces important information more consistently, and self-reports issues more honestly than Opus 5.5, Sonnet 5.5, and Mythos 5.1. Anthropic expects Fable 5.2 to behave similarly.

If a cyber security query is declined, you can retry it on Opus 4.8, and if a biology query is declined, you can retry it on Opus 5. On Amazon Bedrock, the retry is your code’s job, not an automatic switch. If the declined request hit the prompt cache, the refusal also includes a fallback credit token. Send that token with the retry and it’s billed as cache reads, so you don’t pay to write the same cache twice. Check the documentation for more information on fallback credit.

For authorized work in cyber security and life sciences, Claude Mythos 5.2 is available through Anthropic’s verification programs.

Claude Fable 5.2 and Mythos 5.2 use cases

Where Fable 5.2 fits:

  1. Agentic coding across large, multi-repository codebases.
  2. Long-running agents for coordinating many subagents on multi-day work.
  3. Scientific research from new analytical methods to formal mathematical proofs.
  4. Financial analysis for quantitative research and forecasting.
  5. Design for 3D, hardware, and motion for generating environments, components, and simulations in code.

Mythos 5.2 covers two high-risk domains, available only through the verification programs noted earlier:

  1. Cyber security, for security engineering, red teaming, and penetration testing.
  2. Life sciences research, especially biomolecular design and research synthesis.

Data retention and Enterprise Frontier Safeguards

Anthropic designates Claude Fable 5.2 as a Covered Model. By default, its use requires retention for up to 30 days and human safety review. On Amazon Bedrock, Amazon personnel perform the review within the AWS boundary. Set the aws_review data retention mode on your account or project before you invoke the model. See Amazon Bedrock data retention for more details.

Enterprise Frontier Safeguards (EFS), built in collaboration with AWS, combines the privacy of zero data retention (ZDR) with the latest safeguards for detecting misuse. EFS is available now. A version that keeps your data in a bucket and under keys you control is expected to follow, and it has not launched yet. Until it does, eligible customers can use Fable 5.2 with zero data retention through June 30, 2027.

Getting started with Claude Fable 5.2 on Amazon Bedrock

To try Fable 5.2, open the Amazon Bedrock console, choose Test, Playground, and select Fable 5.2 as the model. From there, you can run a prompt directly against it.

Programmatically, you can call the model with the Anthropic Messages API against bedrock-runtime (through the Anthropic SDK). You can also stay on the Invoke and Converse APIs on bedrock-runtime through the AWS Command Line Interface (AWS CLI) and AWS SDK.

Prerequisites

To follow along with the code examples in this post, you need the following:

  1. An active AWS account with Amazon Bedrock access.
  2. AWS CLI installed and configured with appropriate credentials.
  3. Python 3.9 or later.
  4. Boto3 (AWS SDK for Python) installed: pip install boto3.
  5. AWS Identity and Access Management (IAM) permissions: bedrock:InvokeModel and bedrock:InvokeModelWithResponseStream.

Here’s a quick example using the AWS SDK for Python (Boto3):

import boto3 import json # Create a Bedrock Runtime client bedrock_runtime = boto3.client( service_name="bedrock-runtime", region_name="us-east-1" ) # Invoke Claude Fable 5.2 response = bedrock_runtime.invoke_model( modelId="global.anthropic.claude-fable-5-2", contentType="application/json", accept="application/json", body=json.dumps({ "anthropic_version": "bedrock-2023-05-31", "max_tokens": 4096, "messages": [ { "role": "user", "content": "An S3 bucket serves 40 TB/month egress. Estimate the monthly egress cost at $0.09/GB, and state one architecture change, using AWS services, to cut it. Show the calculation, keep it under 120 words." } ] }) ) result = json.loads(response["body"].read()) # Fable 5.2 is a reasoning model: the response may include a thinking block # before the text block, so select the text block rather than a fixed index. print(next(b["text"] for b in result["content"] if b["type"] == "text"))

You can explore the Getting Started notebook for more examples.

Migrating to Claude Fable 5.2 with the claude-api skill in Claude Code

For teams already running Claude models on Amazon Bedrock, the claude-api skill in Claude Code helps you upgrade to Fable 5.2.

The claude-api skill is an open-source skill built into Claude Code that captures the details that make Claude API code work well. It covers which agent patterns fit a given job, what parameters change between model generations, and when to apply prompt caching. It stays current as Anthropic’s SDKs evolve: when a new model is released or the API gains a feature, the skill already knows.

What the skill does for Fable 5.2 migrations

When you ask Claude Code to upgrade your codebase to Fable 5.2, the skill reviews your code and walks you through:

  1. Updating model references swapping model IDs to the Fable 5.2 ID across your application code.
  2. Adjusting thinking settings moving manual thinking configurations to adaptive thinking where appropriate for the new model.
  3. Cleaning up outdated parameters removing deprecated beta headers, parameters, and settings that no longer apply.
  4. Suggesting effort levels recommending the right output_config.effort settings inline based on your use case.

The skill also detects your deployment platform including Amazon Bedrock, Claude Platform on AWS, and other environments and adjusts model ID formats and feature configurations accordingly. For example, it uses the Amazon Bedrock structured model ID format (global.anthropic.claude-fable-5-2).

How to use it

In Claude Code, run the migration directly:

/claude-api migrate

This scans your working directory (or a subdirectory or specific file list you confirm) and applies the migration changes in a single guided pass. You can also invoke it conversationally by telling Claude Code:

  1. “Upgrade me to Claude Fable 5.2” Claude reviews your code and walks through the necessary changes.
  2. “Improve my cache hit rate” the skill applies prompt caching rules that many developers miss, helping reduce costs on Bedrock.
  3. “Add context compaction to my agent” it walks you through compaction primitives and agent patterns for long-running Fable 5.2 sessions.

The claude-api skill is available in Claude Code and is also bundled into CodeRabbit, JetBrains IDEs (including Junie), Resolve AI, and Warp. It is open source at anthropics/skills on GitHub, so any coding agent can integrate it.

Availability

Claude Fable 5.2 and Claude Mythos 5.2 are available today on Amazon Bedrock through US Geo CRIS and Global CRIS inference profiles. See the Bedrock documentation for the full list of supported Regions. For pricing information, see Amazon Bedrock pricing. Both models are also available through Claude Platform on AWS in every Region where the service is offered.

Give Claude Fable 5.2 a try in the Amazon Bedrock console, in Claude Platform on AWS, or explore the Getting Started notebook on GitHub.

About the authors

Dani Mitchell

Dani Mitchell

Dani is a Senior Specialist Solutions Architect for Generative AI at AWS, working on go-to-market for Anthropic on Amazon Bedrock. He helps enterprises across the world design and deploy generative AI solutions using Anthropic’s models and capabilities on Amazon Bedrock to build scalable, production-ready applications.

Anshika Saxena

Anshika Saxena

Anshika is a Generative AI Data Scientist at AWS, focused on Amazon Bedrock. She helps enterprise customers accelerate their generative AI and agentic AI adoption by guiding them through building and implementing solutions on Amazon Bedrock. She also works with ISVs and startups to enable them with the latest Anthropic services on Amazon Bedrock.

Aamna Najmi

Aamna Najmi

Aamna is a Senior Specialist Solutions Architect for Generative AI focusing on Anthropic models and operationalizing and governing generative AI systems at scale on Amazon Bedrock. She helps ISVs solve their challenges, embrace innovation, and create new business opportunities with Amazon Bedrock.

Alfredo Castillo

Alfredo Castillo

Alfredo is a Senior Specialist Solutions Architect for Generative AI at AWS, focusing on Anthropic models go-to-market on Amazon Bedrock. He works with Financial Services customers to design and scale generative AI solutions across distributed systems and turn generative AI experiments into production workloads. Outside of work, he is passionate about family and endurance sports.

Sofian Hamiti

Sofian Hamiti

Sofian is a technology leader with over 12 years of experience building AI solutions, and leading high-performing teams to maximize customer outcomes. He is passionate about empowering diverse talents to drive global impact and achieve their career aspirations.