xAI/grokPublic

Circuit

Code

Versions

API

RUNNING

The circuit is running and is healthy.

0Total Runs

0Avg Response Time

-Avg Tokens per Second

0Avg Spend per Run

-Input Cache Hit Rate

0.00%Error Rate

Control Panel

Explore endpointsModelWorks allows apps to access LLMs and MCP servers worldwide easily in any datacenter, securely and privately.

SecretsOverride default secrets and set up your own for better billing and security.

Circuit Card

Grok X Logo

xAI GROK Chat Completion Service

Name: grok
Author: XAI

Overview

The xAI GROK Chat Completion Service generates AI-driven chat completions using the GROK API. Grok is xAI's flagship AI model with industry-leading capabilities in reasoning, tool calling, and instruction following.

Available Models

Model	Description	Context	Pricing (per 1M tokens)
grok-4.3	Latest flagship model with reasoning capabilities	1M tokens	Input: $1.25 / Output: $2.50
grok-4-1-fast-reasoning	Fast reasoning model with document support	1M tokens	Input: $0.20 / Output: $0.50

Grok 4.3

Key Features

Reasoning: Configurable reasoning effort (none/low/medium/high) with separate reasoning output
Function Calling: Full support for tools and parallel tool calls
Built-in Tools: Real-time web search and X search
Image Understanding: Process images via attachments
1M Context Window: Handle large conversations and documents

Parameters

Required

Parameter	Type	Description
`messages`	array	Conversation messages

Optional

Parameter	Type	Description
`reasoning_effort`	string	Reasoning depth: `none`, `low`, `medium`, `high`
`max_output_tokens`	number	Maximum tokens to generate (includes reasoning tokens)
`temperature`	number	Sampling temperature (0-2)
`top_p`	number	Nucleus sampling value
`tools`	array	Tool definitions (built-in: `web_search`, `x_search`, `code_interpreter`)
`tool_choice`	string	Tool usage: `auto`, `none`, or function name
`parallel_tool_calls`	boolean	Allow parallel tool calls
`attachment`	blob	Image attachment (blob ID or URL)
`user`	string	End-user identifier for abuse monitoring

Outputs

Output	Description
`chat_completion_output`	The generated response text
`chat_reasoning_output`	The reasoning trace (when reasoning enabled)
`chat_completion_tool_calls`	Tool calls made by the model
`chat_completion_finish_reason`	Completion reason: `stop`, `tool_calls`, etc.

Example Requests

Basic Chat

{
  "input": {
    "messages": [
      { "role": "user", "content": "What is 2+2?" }
    ],
    "reasoning_effort": "low",
    "max_output_tokens": 500
  }
}

With Web Search

{
  "input": {
    "messages": [
      { "role": "user", "content": "What are the latest AI news today?" }
    ],
    "tools": [
      {"type": "web_search"},
      {"type": "x_search"}
    ]
  }
}

With Custom Function

{
  "input": {
    "messages": [
      { "role": "user", "content": "Get the weather for San Francisco" }
    ],
    "tools": [
      {
        "type": "function",
        "function": {
          "name": "get_weather",
          "description": "Get weather for a location",
          "parameters": {
            "type": "object",
            "properties": {
              "location": { "type": "string" }
            },
            "required": ["location"]
          }
        }
      }
    ],
    "tool_choice": "auto"
  }
}

Built-in Tools

Tool	Description
`web_search`	Search the web for real-time information
`x_search`	Search X (Twitter) posts, users, threads
`code_interpreter`	Execute Python code in sandbox

The model automatically decides when to use these tools and returns citations in the response.

Cost

Token-based pricing (Grok 4.3): $1.25 per 1M input tokens, $2.50 per 1M output tokens.

Setup

Configure your grok_api_key in your environment secrets
Use the model via the chat completion endpoint

Notes

Use reasoning_effort to control thinking depth - higher values use more tokens but provide better analysis
Enable web_search or x_search in the tools array for real-time data
Pass image blob IDs or URLs via the attachment parameter for image processing

Enjoy dynamic chat completions with xAI GROK!

About

The xAI GROK Chat Completion Service is designed to generate AI-driven chat completions using the GROK API.

About the Author

xAI is an artificial intelligence research company founded by Elon Musk in July 2023, created with the mission to "understand the true nature of the universe" through advanced AI development. The company's flagship product is Grok, a family of large language models designed to compete directly with OpenAI's GPT and Anthropic's Claude. Grok models are deeply integrated with X (formerly Twitter), giving them unique access to real-time social media data and allowing them to provide up-to-date information on

xAI

@xai-org

https://x.ai

@xai

Additional Information

Pricing Structure

The cost model for running this circuit. Charges may be fixed per execution or based on token usage.

Token-based Pricing

UsageLocal

Input tokens0.0₃125 credits

Output tokens0.0₃25 credits

Per Token