# SpaceXAI’s Grok 4.7 now on Amazon Bedrock

> The new model supports longer reasoning and a tunable effort setting for agents, with full accuracy at the highest effort level.

Oossa · 2026-10-13 · https://oossa.com/en/spacexai-s-grok-4-7-now-on-amazon-bedrock

SpaceXAI’s Grok 4.7 is now available through Amazon Bedrock. The model can handle longer, multi‑step tasks and can check its own answers. Bedrock lets you choose a reasoning‑effort level – low, medium, high or xhigh – that controls how much deliberation the model does before responding. In the AWS blog’s tests, the agent got every answer right at xhigh, while using about eight times more tokens than at low.

Grok 4.7 runs behind two inference profiles (us.xai.grok‑4.7 for US‑only data residency, global.xai.grok‑4.7 for wider capacity) and supports the standard OpenAI‑compatible Chat Completions API as well as Bedrock’s Converse API.

## The facts

- Available on Amazon Bedrock as of 2026‑10‑13 ("Tue Oct 13 2026 17:52:20 GMT+0200")
- Correct answers: 18/18 at xhigh effort, 15/18 at low effort

## Why it matters

Developers can tune Grok 4.7’s effort level to trade off cost versus reliability for complex agent workflows.

## Sources & references

1. [Build long-horizon agents with Grok 4.7 on Amazon Bedrock](https://aws.amazon.com/blogs/machine-learning/build-long-horizon-agents-with-grok-4-7-on-amazon-bedrock/) – AWS, 2026-10-13

Last updated: 2026-10-13