Amazon Bedrock and Strands Updates for September 2026
Key facts
- Amazon Bedrock Managed Agents powered by OpenAI are now available in public preview.
- The latest AgentCore runtime features more efficient memory management and lower cold start latency.
- Strands harness is an open source agent harness that matches popular harnesses on accuracy while using 28 percent fewer tokens.
- Strands Decider 2B is an open source 2B-parameter decision model that picks between predefined options and delivers answers in about 115ms locally.
- OpenAI Astra, Sol, and Luna models are now generally available on Amazon Bedrock.
Amazon Bedrock Managed Agents
Amazon Bedrock Managed Agents powered by OpenAI are now available in public preview. This allows builders to construct production-ready AI agents while maintaining enterprise security.
Users can use OpenAI models while keeping data within AWS. The system lets teams reuse existing AWS Identity and Access Management permissions and maintain auditability through AWS CloudTrail.
Durable sessions and built-in human approval workflows assist in reducing development overhead and mitigating risk. Agents start faster with less infrastructure overhead.
AgentCore Runtime Updates
The latest AgentCore runtime introduces more efficient memory management. It also provides lower cold start latency for serverless agents.
Users pay for actual usage rather than peak memory. Agents run faster at lower cost with pay-as-you-go pricing without needing to pre-provision capacity.
Sessions scale to zero when idle. They run inside hardware-isolated environments.
Strands Toolkit Enhancements
The Strands harness is a new open source agent harness. It matches popular harnesses on accuracy while utilizing 28 percent fewer tokens.
Production-ready agents can be spun up in a single line of Python or TypeScript. It includes built-in context management, prompt caching, and memory for deployment anywhere.
Strands Decider 2B is a small, open source, 2B-parameter decision model. It picks between predefined options instead of generating text and delivers answers locally in about 115ms.
OpenAI Models on Amazon Bedrock
OpenAI models including Astra, Sol, and Luna are now generally available on Amazon Bedrock. They offer options for demanding projects, recurring complex work, and high-volume tasks.
GPT-6 Astra serves as the flagship model for ambitious work. It brings deeper reasoning to complex decisions, document analysis, and software development with up to 1 million input tokens.
GPT-6 Astra Ultrafast adds a premium speed tier for workloads requiring high speed. It delivers up to 6x faster inference in the API with up to 300 tokens per second.
Claude and Additional Model Additions
Model choices for Claude now feature newer versions for coding and scientific work. Claude Fable 5.1 expands options for coding, scientific research, and enterprise workflows.
Claude Opus 5.5 is designed for agentic coding, knowledge work, and long-running tasks. It uses adaptive thinking and an effort parameter to set an upper bound on reasoning depth.
Moonshot AI’s Kimi K3 is available on Amazon Bedrock. According to Moonshot AI, Kimi K3 is its most capable model and the first open model to reach 2.8 trillion parameters.
Enterprise Knowledge Base Updates
Automatic sync scheduling in Amazon Bedrock Managed Knowledge Base enables daily, weekly, and monthly refresh options. This applies to all native data-source connectors to keep enterprise agent knowledge current.
User-managed setups for SharePoint, OneDrive, and Confluence reduce reliance on admin-managed service accounts. Users can connect enterprise sources without maintaining custom ingestion pipelines.
Frequently asked questions
What models are included in the OpenAI updates for Amazon Bedrock?
OpenAI models including Astra, Sol, and Luna are generally available on Amazon Bedrock. Specific versions include GPT-6 Astra, GPT-6 Astra Ultrafast, GPT-6.1 Sol, GPT-6 Sol, and GPT-6.1 Luna.
What is Strands Decider 2B?
Strands Decider 2B is a small, open source, 2B-parameter decision model that chooses between predefined options instead of generating text. It delivers answers locally in roughly 115 milliseconds.
What new native connectors are supported by Amazon Bedrock Managed Knowledge Base?
Amazon Bedrock Managed Knowledge Base now supports ServiceNow, Confluence Data Center, Salesforce, and Zendesk as native connectors.
How does automatic sync scheduling work in Amazon Bedrock Managed Knowledge Base?
Automatic sync scheduling lets users configure daily, weekly, and monthly refresh options for all native data-source connectors.