AWS Public Sector Blog
Amazon Bedrock launches with Claude 3.5 Sonnet in the AWS Top Secret cloud
Amazon Bedrock is now generally available in the Amazon Web Services (AWS) Top Secret cloud. With authority to operate in the AWS Top Secret cloud, Amazon Bedrock is providing U.S. Government customers with secure access to foundation models (FMs) and generative artificial intelligence (AI) capabilities within AWS’s classified environment. Amazon Bedrock enables customers to leverage generative AI capabilities while maintaining the highest levels of security and data protection required for their Top Secret workloads.
“This milestone enables us to bring the power of frontier models like Anthropic’s upgraded Claude 3.5 Sonnet to our U.S. Government customers,” said Dave Levy, Vice President of Worldwide Public Sector at AWS. “We are delivering capabilities that will reshape how national security organizations process information, make decisions, and support critical missions.”
What is Amazon Bedrock?
Amazon Bedrock is a scalable, fully managed service that offers a choice of high-performing FMs from leading AI companies through a single API, along with a broad set of capabilities to build generative AI applications with security, privacy, and responsibility. Using Amazon Bedrock, customers can easily experiment with and evaluate frontier multimodal models for their use cases and connect models to their data using techniques such as Retrieval-Augmented Generation (RAG). Since Amazon Bedrock is serverless, customers don’t have to manage any infrastructure, and they can securely integrate and deploy generative AI capabilities into their applications using the AWS services they are already familiar with.
Model and feature availability
Amazon Bedrock launched in the AWS Top Secret cloud with Anthropic’s upgraded Claude 3.5 Sonnet and the Amazon Titan Text Embedding v2 models. Both models are accessible on demand.
- Anthropic’s upgraded Claude 3.5 Sonnet. Customers have access to the latest Claude multimodal model that delivers across-the-board improvements over its predecessor, with particularly significant gains in coding—an area where it already led the field, according to Anthropic. The model also raises the bar across a range of cognitive tasks. It more deeply understands nuanced instructions and context, generates creative solutions to problems, recognizes and corrects mistakes, and delivers more sophisticated analysis around complex data. Other use cases in which the upgraded model excels include knowledge Q&A platforms, data extraction from visuals like charts and diagrams, and automation of repetitive tasks and operations.
- Amazon Titan Text Embedding v2. Using Titan Text Embeddings, customers can perform various natural language processing (NLP) tasks by representing text data as numerical vectors, known as embeddings. These embeddings capture the semantic and contextual relationships between words, phrases, or documents in a high-dimensional vector space. This model is optimized for RAG use cases and is also well suited for a variety of other tasks such as information retrieval, Q&A chatbots, classification, and personalized recommendations.
Amazon Bedrock in the AWS Top Secret cloud features the console, playgrounds, and Guardrails. Knowledge Bases and Agents are planned for future releases, along with additional FMs.
Use cases for defense and intelligence
Amazon Bedrock offers powerful AI capabilities tailored for defense and intelligence applications.
Writing assistant
Streamline the creation of intelligence reports, operational briefs, and strategic communications. Amazon Bedrock’s writing assistance can help analysts and operators produce clear, concise documents while adhering to specific security protocols and terminology.
Coding assistant
Accelerate the development of mission-critical software and data analysis tools. Anthropic’s Claude coding capabilities can assist software engineers in writing secure, efficient code for defense systems and intelligence platforms.
Intelligent search and analysis with RAG
Enhance AI responses with classified and sensitive data repositories. RAG allows intelligence agencies to combine large language models (LLMs) with their mission databases, enabling more accurate and contextually relevant analysis for complex geopolitical scenarios and threat assessments.
These capabilities empower customers working with classified data to process information more efficiently, make data-driven decisions faster, and stay ahead in an ever-evolving global security landscape.
Why use Amazon Bedrock?
There are three compelling reasons to use Amazon Bedrock for building generative AI applications:
Access to leading FMs
Amazon Bedrock offers an easy-to-use developer experience to work with Anthropic’s Claude multimodal model and Amazon Titan Text Embedding model. Customers can quickly experiment in the playground, and use a single API for inference to keep up to date with the latest model versions with minimal code changes.
Extend the power of FMs with proprietary data
Customers can securely connect FMs to their data sources for retrieval augmentation, extending the FM’s already powerful capabilities and making it more knowledgeable about customer-specific domain and organization. Native support for RAG through Amazon Bedrock Knowledge Bases will be available in the future.
Data security and compliance certifications
Amazon Bedrock offers several capabilities to support security and privacy requirements. With Amazon Bedrock, customer content and queries are not retained, are not used to improve the base models, and are not shared with any model providers. Customer data in Amazon Bedrock is encrypted in transit and at rest, and customers can optionally encrypt the data using their own keys. Customers can use AWS PrivateLink with Amazon Bedrock to establish private connectivity between their FMs and their Amazon Virtual Private Cloud (Amazon VPC) without exposing their traffic to any shared networks in or out of AWS.
What’s next in the AWS Top Secret Cloud
Throughout 2025, AWS will continue to launch generative AI capabilities in the AWS Top Secret cloud and other classified regions.
“Our mission is clear: to empower our U.S. Government customers with cutting-edge generational capabilities,” said Dave Nolton, Vice President of Region Services at AWS. “We are focused on delivering the right models, services, and capabilities across our hyper scale classified infrastructure. The General Availability of Amazon Bedrock in the AWS Top Secret cloud will enable choice for these customers. As we have across the history of our partnership, we continue to innovate in support of our customers.”
We are excited to bring Amazon Bedrock to our U.S. Government customers. If you have questions or feedback, please contact your AWS Account Team.