AWS AI Updates: June 10, 2026

1. AWS Makes Claude Fable 5, the First Generally Available Mythos-Class Model, Available on Bedrock

AWS. AWS announced the general availability of Claude Fable 5, which it describes as the first generally available model in Anthropic’s Mythos-class capability tier, targeting extended autonomous work on complex knowledge and coding tasks with abilities such as self-updating skills, building evaluation frameworks, and independently verifying its own outputs. The model is offered through two paths: Amazon Bedrock, which keeps data within AWS infrastructure and supports managed features like Guardrails and Knowledge Bases, and the Claude Platform on AWS, operated by Anthropic with unified billing. A separate Claude Mythos 5 variant without safety classifiers is restricted to preview customers, while the general release ships with safeguards intended for broader production use. Source

2. AWS FinOps Agent Enters Preview for Conversational Cloud Cost Management

AWS. AWS launched the FinOps Agent in preview, an AI-powered agent that lets finance and engineering teams ask natural-language questions about AWS spending, generate financial reports, and surface optimization opportunities such as rightsizing and idle resource detection. The agent can automatically investigate cost anomalies and their root causes, create Jira tickets, and post findings to Slack channels, and it is accessed through the AWS Management Console. During preview it is available in US East (N. Virginia) with cost data coverage spanning all AWS Regions except GovCloud (US) and China, and there are no additional charges for the preview. Source

3. SageMaker Unified Studio Notebooks Add EMR Serverless as a Spark Runtime

AWS. Amazon SageMaker Unified Studio Notebooks now let users run PySpark and Spark SQL on an EMR Serverless Spark application directly from notebook cells, adding a runtime option alongside the existing Athena Spark support and selectable from the notebook side panel. The integration uses pre-initialized capacity to reduce session start times, provides a consistent Spark UI for job monitoring across engines, and supports VPC connectivity for network isolation, with the SageMaker Data Agent able to generate Spark code from natural-language prompts. The feature is available in all AWS Regions where SageMaker Unified Studio is offered and works in both notebook and JupyterLab IDE environments. Source

4. Cost Explorer Adds Amazon Q-Powered Intelligent Cost Explanations

AWS. AWS added an “Analyze with Amazon Q” capability to Cost Explorer that uses Amazon Q Developer to automatically explain cost trends, identify top cost drivers, and flag anomalies based on the exact filters and time period selected in the console view. The analysis adapts to historical, forecasted, or mixed time ranges, returns results in a chat interface, and supports follow-up questions that retain conversation context for deeper exploration. It is available in all commercial AWS Regions at no additional charge through the Cost Explorer console. Source