ICYMI: What landed for AI builders in October 2026
A recap of the major Amazon Bedrock, Amazon Bedrock AgentCore, and Strands updates from October 2026: new models to match each task, updates that keep agents available and grounded, and clearer cost reporting for AI builders.
In October 2026, Amazon Web Services (AWS) introduced a range of updates aimed at supporting AI builders. The company's Amazon Bedrock service provides access to leading models and tools for building and scaling AI applications. A key addition is Amazon Bedrock AgentCore, a platform for building, connecting, and optimizing agents using any model and framework.
One highlight is the launch of Claude Haiku 5.5, Anthropic's most capable Haiku model yet, which costs around 75% less than its predecessor for most tasks. It also features effort controls, allowing users to balance cost and intelligence for each task. When used alongside other models, it enables efficient parallel processing in multi-agent setups.
Meanwhile, OpenAI introduced GPT-6.1 Terra, which matches the performance of GPT-6 Astra while completing tasks more than three times faster. It also offers an Ultrafast mode for latency-sensitive work, such as real-time coding assistants. For voice agents, Amazon Nova 2.5 Sonic improves reasoning, instruction following, and tool-calling accuracy with lower latency.
Other additions include GLM 5.3 by Z.ai, an open-weight model for coding and long-horizon software engineering, and TwelveLabs Pegasus 1.5, which converts videos into structured, time-coded metadata. Additionally, AWS GovCloud (US) now offers Claude Haiku 5.5, GPT-5.6, Daybreak Red, and cross-Region inference capabilities for security work.
AWS Step Functions workflows can now be used as tools through AgentCore Gateway, integrating existing business logic seamlessly. Lastly, structured JSON outputs are now available for all Claude models on Amazon Bedrock, enabling direct use of the data in other systems.
Written by urgent.news from AWS Machine Learning's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
