Grok 4.7
The current Grok flagship accepts text and images and returns text. It exposes reasoning effort, tool calls and structured output.
What the release provides
Effort settings run from low to xhigh, with high as the default. Standard pricing changes at 200,000 input tokens. Batch access is listed, with a separate discount.Grok 4.7 API model card
How to use it well
Practical guidance. For a research assistant, separate source retrieval from synthesis. Request citations from the retrieved material and check whether the answer actually uses them. Compare effort levels on tasks with a checkable final result before choosing a default.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API |
| Model identifiers | grok-4.7 |
| Context window | 500,000 tokens Published API context limit Grok 4.7 API model card |
| Input price | 2 USD per million tokens Standard uncached input; prompt below 200,000 tokens; higher rate applies to all tokens at the threshold Grok 4.7 API model card |
| Output price | 6 USD per million tokens Standard output; prompt below 200,000 tokens; excludes tools and batch discounts Grok 4.7 API model card |