Amazon Bedrock adds xAI's Grok 4.6 with 500K token context window
AWS said the model runs on two Bedrock endpoints, with four reasoning effort levels and support for cross-Region inference.
xAI's Grok 4.6 is now available in Amazon Bedrock, Amazon Web Services said in a post on Monday. The company put the context window at 500,000 tokens and described the model as built for long-running agents and for coding.
The characterisation is xAI's and Amazon's own, offered in a vendor post. AWS published no benchmark scores with the announcement, and no independent evaluation of Grok 4.6 running on Bedrock has appeared in the material available so far.
The model runs on both the bedrock-mantle and bedrock-runtime endpoints, AWS said, with Converse API and cross-Region inference support. The post lists four reasoning effort levels, and does not say what separates them. AWS gave no price for the model in the post.
Grok 4.6 is the version before Grok 4.7. The account theo, posting on X, said Grok 4.7 does not beat Grok 4.6 on many benchmarks, and called Grok 4.6 slower and more expensive than Grok 4.5. Those are one user's assessments, and may not hold across other tasks.