Grok 4.7 is now available on Amazon Bedrock
xAI’s Grok 4.7, now on Bedrock, can handle 500K‑token contexts, self‑verify outputs, and run long coding/knowledge tasks better than before.
Why Now
xAI released Grok 4.7 on September 21 2026 and announced it on Bedrock the next day, adding a new frontier model for coding and long‑running agents.
What Happened
Grok 4.7 offers a 500K‑token context window, four configurable reasoning levels, and supports Responses, Chat Completions, and Converse APIs. It outperforms Grok 4.6 on independent benchmarks, especially in long‑horizon agentic work and coding, but produces roughly double the output tokens per task. The model includes a new safeguard stack and is reported to be the strongest on refusals and jailbreak resistance.
Why It Matters
The larger context and self‑verification make Grok 4.7 suitable for building reliable agents that need to reason over long documents or codebases. The higher token output means higher compute costs, so users must tune reasoning levels. Strong safety features lower the risk of misuse in dual‑use domains like cybersecurity.
The Limitation
The token‑cost increase and need for careful reasoning‑level tuning may offset the performance gains for some workloads.
What You Can Do
Try Grok 4.7 on Bedrock with a custom reasoning level and benchmark its token usage on your long‑form task.
Source
Read original sourceWhy we picked this
New xAI model on Bedrock with large context and APIs – core AI content.