AWS Weekly Roundup: Claude Mythos Preview in Amazon Bedrock, AWS Agent Registry, and more (April 13, 2026)

AWS continues to expand its AI and cloud infrastructure capabilities with several significant announcements this week, addressing critical needs around cost visibility, security, and operational efficiency. From enhanced Bedrock features to new quantum computing access, here's what caught my attention.

Enhanced Cost Visibility for AI Workloads

One of the most impactful announcements this week addresses a growing pain point I've observed in AI-DLC workshops: cost allocation. As organizations scale their AI initiatives, finance teams need clear visibility into who's using what resources and at what cost.

Amazon Bedrock now supports cost allocation by IAM user and role, a feature that fundamentally changes how organizations can track and manage their AI investments. This capability allows teams to tag IAM principals with attributes like team or cost center, then activate those tags in the Billing and Cost Management console. The resulting cost data flows into AWS Cost Explorer and the detailed Cost and Usage Report.

This is particularly valuable for organizations running tools like Claude Code on Amazon Bedrock or scaling agents across multiple teams. The ability to track foundation model usage by department provides the financial transparency needed to justify AI investments and optimize spending patterns.

Claude Mythos: Anthropic's Cybersecurity-Focused AI Model

Anthropic's most sophisticated AI model to date, Claude Mythos, is now available on Amazon Bedrock as a gated research preview through Project Glasswing. This represents a significant evolution in AI capabilities, introducing a new model class specifically focused on cybersecurity.

Claude Mythos demonstrates state-of-the-art performance across cybersecurity, coding, and complex reasoning tasks. Security teams can leverage this model to discover and address vulnerabilities in critical software before threats emerge. The model's ability to identify sophisticated security vulnerabilities in software and analyze large codebases makes it particularly valuable for organizations with complex security requirements.

Access is currently limited to allowlisted organizations, with Anthropic and AWS prioritizing internet critical companies and open source maintainers. This selective rollout approach ensures the model's powerful capabilities are deployed responsibly in environments where they can have the greatest impact.

AWS Agent Registry: Centralized AI Agent Management

AWS launched Agent Registry through Amazon Bedrock AgentCore, providing organizations with a private catalog for discovering and managing AI agents, tools, skills, MCP servers, and custom resources. This addresses a common challenge I've seen in enterprise AI deployments: teams often duplicate capabilities because they don't know what's already available.

The registry offers semantic and keyword search capabilities, approval workflows, and CloudTrail audit trails. Teams can access it via the AgentCore Console, AWS CLI, SDK, and as an MCP server queryable from IDEs. This centralized approach to agent discovery and governance helps organizations avoid the inefficiencies of siloed AI development while maintaining proper oversight and compliance.

Major Infrastructure Updates

Amazon S3 Files: File System Semantics for S3

Amazon S3 Files transforms S3 buckets into shared file systems that connect any AWS compute resource directly with S3 data. Built on Amazon EFS technology, it delivers full file system semantics with low latency performance, caching actively used data and providing multiple terabytes per second of aggregate read throughput.

Applications can access S3 data through both file system and S3 APIs simultaneously without code modifications or data migration. This eliminates the traditional trade-off between object storage scalability and file system usability, making it easier for applications to leverage S3's durability and cost advantages while maintaining familiar file system interfaces.

Unified Observability with OpenSearch

Amazon OpenSearch Service now provides a unified observability platform that consolidates metrics, logs, traces, and AI agent tracing into a single interface. The update includes native Prometheus integration with direct PromQL query support, RED metrics monitoring, and OpenTelemetry GenAI semantic convention support for LLM execution visibility.

Operations teams can correlate slow traces to logs and overlay Prometheus metrics on dashboards without switching between tools. This unified approach to observability significantly reduces the complexity of monitoring modern, distributed applications that often span multiple services and data sources.

AI-Powered WorkSpaces Troubleshooting

AWS launched Amazon WorkSpaces Advisor, an AI-powered administrative tool that uses generative AI to help IT administrators troubleshoot Amazon WorkSpaces Personal deployments. It analyzes WorkSpace configurations, detects problems automatically, and provides actionable recommendations to restore service and optimize performance.

This represents a practical application of AI to IT operations, where the complexity of modern desktop environments often makes troubleshooting time-consuming and error-prone. The AI-powered approach can identify patterns and correlations that might be missed by human administrators, leading to faster resolution times and improved user experiences.

Quantum Computing Advancement

Amazon Braket now offers access to Rigetti's Cepheus-1-108Q device, the first 100+ qubit superconducting quantum processor on the platform. The modular design features twelve 9 qubit chiplets with CZ gates that offer enhanced resilience to phase errors.

The device supports multiple frameworks including Braket SDK, Qiskit, CUDA-Q, and Pennylane, with pulse level control for researchers. This expansion of quantum computing capabilities provides researchers and organizations with access to more powerful quantum processors, accelerating the development of quantum algorithms and applications.

Looking Ahead

These announcements reflect AWS's continued investment in making AI and cloud infrastructure more accessible, manageable, and powerful. The focus on cost visibility, security, and operational efficiency addresses real-world challenges that organizations face as they scale their cloud and AI initiatives.

For organizations planning their AI strategies, these updates provide new tools for managing costs, enhancing security, and improving operational efficiency. The combination of Claude Mythos for advanced security analysis, Agent Registry for centralized governance, and enhanced observability capabilities creates a more robust foundation for enterprise AI deployments.

The upcoming "What's Next with AWS" event on April 28th promises to provide further insights into how agentic AI is transforming business operations, featuring AWS leadership and OpenAI executives discussing emerging capabilities and platform innovations.

As AI continues to evolve from experimental projects to core business capabilities, these infrastructure improvements will play a crucial role in enabling organizations to scale their AI initiatives responsibly and effectively.