Grok 4.7 from xAI available on Amazon Bedrock
The company xAI has introduced Grok 4.7 on Amazon Bedrock: a context window of 500 000 tokens and four reasoning effort levels. According to Artificial Analysis, performance on agentic tasks and knowledge work improves compared with Grok 4.6, but at roughly twice the token consumption.
The company xAI has made Grok 4.7 available on Amazon Bedrock, expanding the catalog of models available through this service from Amazon. The model is designed for programming, long-running agentic tasks and knowledge work, offers a context window of 500 000 tokens and supports four configurable reasoning effort levels: low, medium, high and xhigh. The model runs through the bedrock-runtime endpoint with cross-Region inference profiles and supports Responses API, Chat Completions API, InvokeModel and Converse API.
According to xAI, this is their most capable model for coding and knowledge work, built on a new and larger base model with longer reinforcement learning training on more difficult tasks specifically focused on problems that take multiple hours. The company says this enables the model to verify its own outputs better and use long context more effectively on longer-running tasks. xAI further claims that the model delivers improvements in document and presentation generation and in professional knowledge work (law, healthcare, financial analysis), and that it has been equipped with a new safety framework with stronger resistance to jailbreaks.
Independent measurements from Artificial Analysis show improvements over the previous model, Grok 4.6: Intelligence Index rose from 44 to 46, Coding Agent Index from 47 to 56, the AA-Briefcase score (long-term knowledge work, Elo) from 1546 to 1657, GDPval-AA (professional outputs, Elo) from 1605 to 1695 and AA-Omniscience Index from 30 to 32, while the hallucination rate fell from 34 % to 29 %. However, these improvements come with nearly twice the output token consumption per task (approximately 81 thousand compared with 38 thousand), with Grok 4.7 measured at the highest level, xhigh.
Details on technical integration with Bedrock, pricing and additional safety measures can be found in the source article.
Why it matters
For developers and companies working with Amazon Bedrock, this is another model option for agentic and coding tasks with long context that, according to independent measurements from Artificial Analysis, improves reliability on long-running tasks and reduces the hallucination rate. However, the measured token consumption is a key finding — the highest reasoning effort level delivers better results at nearly twice the output cost, so the choice of reasoning effort level directly affects deployment operating costs.
Two audiences, two different impacts
What this means
For individuals
Developers working with Amazon Bedrock have another model available for coding and long-running agentic tasks. According to the data, it is worth deliberately choosing the reasoning effort level because of the noticeable difference in token consumption and therefore in response speed and cost.
For a business
Companies building agentic systems or automation on Amazon Bedrock gain another frontier model with a long context window and self-verification that, according to Artificial Analysis, reduces error rates on long tasks but consumes roughly twice as many output tokens as the previous version at the highest reasoning effort level, increasing operating costs.
DevelopmentCheck the original
Event sources
clearly official source · 1 publisher, 0 independent. We count feeds from the same owner only once.