Amazon.com Inc.

11/09/2026 | News release | Distributed by Public on 11/09/2026 14:34

Claude Fable 5.2 is now available on AWS

Key takeaways

  • Claude Fable 5.2 sets a new standard on complex technical work.
  • It runs for hours to days unattended, coordinating subagents and verifying its own output.
  • Claude Mythos 5.2 is available through verification programs for cybersecurity and life sciences.
Claude Fable 5.2 and Claude Mythos 5.2 are now available on Amazon Bedrock and Claude Platform on AWS. According to Anthropic, Fable 5.2 is its most capable generally available model for technical work. Mythos 5.2 is available through Anthropic's verification programs for advanced work in cybersecurity and life sciences.
In Anthropic's alignment testing, Mythos 5.2 (the same underlying model as Fable 5.2, without safeguards) scored lower than Opus 5.5, Sonnet 5.5, and Mythos 5.1 on behaviors such as hallucinating inputs and falsely claiming to have completed a task. Anthropic expects Fable 5.2 to behave similarly. Safeguards limit misuse in sensitive domains such as cybersecurity, life sciences, and model distillation. Where a cybersecurity or biology query is blocked, customers instead receive a response from Opus 4.8 for cybersecurity and Opus 5 for biology. For authorized work in cybersecurity and life sciences, Claude Mythos 5.2 is available through Anthropic's verification programs.
Fable 5.2 runs for hours to days unattended-interpreting requirements from high-level prompts, iterating on its own output, flagging inconsistencies during self-verification, and driving performance optimizations. In Anthropic's testing, it ran multi-day tasks and coordinated hundreds or even more than a thousand subagents, building controls and flagging results that don't reproduce.
The new model handles agentic coding across repositories, scientific research including finding new methods to complete hard physics computations or formalize very large mathematical proofs, financial analysis from multi-agent research to forecasting, and data analysis.

AWS commits $100 million of AI and cloud technology to help organizations expand health services worldwide

The AWS Health Impact Initiative gives health organizations access to AI and cloud technology to build solutions for underserved communities.

Claude Fable 5.2 shows improvement in the following areas

According to Anthropic, Fable 5.2 sets a new standard on the hardest projects in coding, math, and science. Here's what sets it apart:
  1. Frontier-level technical performance. Handles the hardest projects in coding, math, and science, verifying its own output and refining results iteratively.
  2. Stronger safety profile. In Anthropic's alignment testing, Mythos 5.2 (the same underlying model as Fable 5.2, without Fable 5.2's cyber and bio safeguards) scored lower than Opus 5.5, Sonnet 5.5, and Mythos 5.1 on behaviors such as hallucinating inputs and falsely claiming to have completed a task. Anthropic expects Fable 5.2 to behave similarly. Safeguards limit misuse in sensitive domains such as cybersecurity, life sciences, and model distillation. For authorized work in cybersecurity and life sciences, Claude Mythos 5.2 is available through Anthropic's verification programs.
  3. Autonomous, long-running operation. Runs for hours to days unattended. It interprets requirements from high-level prompts, iterates on its own output, flags inconsistencies during self-verification, and drives performance optimizations.
  4. Agentic coding. Takes on loosely defined engineering projects across repositories and monorepos. It applies architectural judgment, tests its work incrementally, and questions inherited assumptions to surface root causes earlier Claude models miss.
  5. Long-running agents and orchestration. Acts as lead agent on open-ended work. In Anthropic's testing, it ran multi-day tasks and coordinated hundreds or even more than a thousand subagents, building controls and flagging results that don't reproduce.
  6. Scientific research. Develops new methods for physics computations, engineers fully functional electronic devices, and formalizes mathematical proofs.
  7. Financial analysis. Handles quantitative finance, from multi-agent research to forecasting. It catches subtle edge cases and builds reports that could take several days of effort.

How AWS is helping companies build physical AI machines that think

Amazon launched the Physical AI Toolchain on AWS, an open-source stack that helps manufacturers build machines that perceive, reason, and act in the real world.

How to access Claude Haiku 5.5 on AWS

Customers can get started with Claude Fable 5.2 through Amazon Bedrock or Claude Platform on AWS:
  • Amazon Bedrock: Build within your existing AWS environment and scale inference workloads. Access the model programmatically using the Anthropic Messages API, the Invoke and Converse APIs via the AWS CLI and AWS SDK, or use it in the Bedrock console Playground.
  • Claude Platform on AWS: Access through Anthropic's native platform experience, available in North America, South America, Europe, and Asia Pacific.

Dive deeper

Visit Anthropic's Claude in Amazon Bedrock product page for more detailed information on Claude Fable 5.2, Claude Haiku 5.5, Claude Sonnet 5.5, Claude Opus 5.5, Claude Fable 5.1, Claude Opus 5, Claude Sonnet 5, Claude Fable 5, Claude Opus 4.8, Claude Opus 4.7, Claude Sonnet 4.6, Claude Opus 4.6, Claude Haiku 4.5, Claude Sonnet 4.5, Claude Opus 4.5, Claude Opus 4.1, and Claude Sonnet 4.
Update: This article was originally published on June 9, 2026. It has been updated on November 9, 2026 to reflect the availability of Claude Fable 5.2 and Claude Mythos 5.2.
October 07, 2026: This article has been updated to reflect the availability of Claude Haiku 5.5.
September 28, 2026: This article has been updated to reflect the availability of Claude Sonnet 5.5.
September 22, 2026: This article has been updated to reflect the availability of Claude Opus 5.5.
September 1, 2026: This article has been updated to reflect the availability of Claude Fable 5.1.
July 24, 2026: This article has been updated to reflect the availability of Claude Opus 5.
June 30, 2026: This article has been updated to reflect the availability of Claude Sonnet 5.
June 9, 2026: This article has been updated to reflect the availability of Claude Fable 5.
May 28, 2026: This article has been updated to reflect the availability of Claude Opus 4.8.
February 17, 2026: This article has been updated to reflect the availability of Claude Sonnet 4.6.
February 5, 2026: This article has been updated to reflect the availability of Claude Opus 4.6.
November 24, 2025: This article has been updated to reflect the availability of Claude Opus 4.5.
October 15, 2025: This article has been updated to reflect the availability of Claude Haiku 4.5.
September 29, 2025: This article has been updated to reflect the availability of Claude Sonnet 4.5.
August 5, 2025: This article has been updated to reflect the availability of Claude Opus 4.1.
Amazon.com Inc. published this content on November 09, 2026, and is solely responsible for the information contained herein. Distributed via Public Technologies (PUBT), unedited and unaltered, on November 09, 2026 at 20:34 UTC. If you believe the information included in the content is inaccurate or outdated and requires editing or removal, please contact us at [email protected]