Claude Fable 5 Now Available on Amazon Bedrock
Anthropic has launched Claude Fable 5 on Amazon Bedrock and the Claude Platform on AWS, introducing long-running asynchronous execution and advanced vision capabilities. According to the company, the model features proactive self-verification and a tiered safety system that routes high-risk queries to a secondary model, Opus 4.8, to prevent misuse in sensitive domains.
Why is asynchronous execution a trend for AI agents?
Claude Fable 5 shifts AI interaction from a turn-based chat to long-running, asynchronous execution. According to Anthropic, the model can now execute coding and knowledge work tasks for extended periods without human intervention.
This capability suggests a move toward autonomous agents that handle complex, multi-step workflows. Rather than providing a single answer, the model sustains a task until completion. This is particularly relevant for software engineering, where the model can build, test, and refine code independently.
bedrock-mantle or bedrock-runtime endpoints.
How will advanced vision change document-heavy industries?
The model’s ability to understand diagrams, charts, and tables nested within PDFs targets specific professional sectors. Anthropic states this opens up new research and document-heavy workflows in finance, legal, analytics, architecture, and gaming.
In software development, the model uses vision to critique its own output against design goals. This feedback loop allows for higher fidelity in implementing designs. The trend points toward AI that does not just generate text, but verifies its visual accuracy against a source blueprint.
What does the split between Fable and Mythos mean for AI safety?
Anthropic has introduced a tiered capability structure to manage risk. Claude Fable 5 includes safeguards that limit performance in areas like cybersecurity, biology, chemistry, and health. When a prompt triggers these safeguards, the system falls back to Opus 4.8.
For high-stakes domains, the company offers Claude Mythos 5. This unrestricted version is available only to a small group of vetted customers. According to the documentation, Mythos 5 is designed for vulnerability discovery, drug design, and biodefense screening.
How is data retention evolving for high-capability models?
Access to Fable 5 and Mythos 5 requires a specific data-sharing agreement. Users must opt into data retention using the Data Retention API and setting provider_data_share.
Anthropic requires 30-day retention for all inputs and outputs, including human review. The company states this is necessary to detect patterns of misuse that are not visible in single exchanges. This requirement means data leaves the AWS security boundary, marking a shift in how privacy is balanced with safety for state-of-the-art models.
Comparison: Fable 5 vs. Mythos 5
| Feature | Claude Fable 5 | Claude Mythos 5 |
|---|---|---|
| Availability | Broad AWS accounts | Vetted customers only |
| Safety Limits | Active (Fallbacks to Opus 4.8) | Unrestricted for specific use cases |
| Primary Use | Knowledge work & Engineering | Cybersecurity & Life Sciences |
Frequently Asked Questions
Where is Claude Fable 5 available?
It is available on Amazon Bedrock in the US East (N. Virginia) and Europe (Stockholm) regions, and on the Claude Platform on AWS across North America, South America, Europe, and Asia Pacific.
What is the data retention requirement for Fable 5?
Anthropic requires a 30-day retention period for all traffic on Mythos-class models to detect misuse patterns.
How do I enable Claude Fable 5 on AWS?
Users must programmatically opt into data sharing via the Data Retention API. There is currently no console user interface for this setting.
Do you think asynchronous AI agents will replace traditional coding workflows? Share your thoughts in the comments below or subscribe to our newsletter for more AI industry updates.