xAI’s Grok 4.6 is now available in Amazon Bedrock

xAI has made its Grok 4.6 model available on Amazon Bedrock, expanding the frontier model's capabilities for long-running agents, coding, and knowledge work. Grok 4.6 launched on Bedrock on August 18, 2026, and provides a 500K token context window, along with four reasoning effort levels: low, medium, high, and xhigh. This is xAI's second model in Amazon Bedrock.

Grok 4.6 enhances the existing surface area by being available on both the bedrock-mantle and bedrock-runtime endpoints, supporting the Converse API alongside Chat Completions and Responses. The model is designed for complex tasks that span multiple steps, such as researching a topic, analyzing information, working across a codebase, or turning an idea into a polished application or work artifact.

xAI trained Grok 4.6 using a longer supplemental training run, with a focus on reasonings and advanced technical concepts, high-quality engineering data, and improved optimizer and training recipe. The model was then trained on various agentic reinforcement learning tasks, including knowledge work, general coding, and domain-specific environments such as kernel optimization, web development, and computer-aided design.

Several safety improvements have been made to Grok 4.6, including better safeguards that have been calibrated in line with the model's capabilities. The model has undergone a wide range of pre-deployment testing for capabilities and safeguard calibration, as well as post-deployment and third-party testing. This safety stack aims to maximize utility and security across legitimate use cases in areas like vulnerability patching, accelerating the engineering design cycle, and augmenting AI research.