How much does the Grok 4.6 API cost?

On FastMetal, Grok 4.6 is billed per token: $2.20 per 1M input tokens and $6.60 per 1M output tokens on a US-dollar account, before tax. Usage is drawn from a prepaid balance; there is no subscription or monthly fee.

Can I call Grok 4.6 with the OpenAI SDK?

Yes. Point base_url at https://api.fastmetal.ai/v1 and pass "grok-4.6" as the model; existing OpenAI-style code works unchanged, including streaming, tool calls and structured output.

What is the context window of Grok 4.6?

500,000 tokens, shared between the prompt and the response.

What do I need to try Grok 4.6?

Create an account and add credit; the model is then available both in the browser chat and over the API. There is no contract or minimum spend.

Back to Models

xAI

grok-4.6

Grok 4.6

Grok 4.6 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.

8/12/2026

500,000 tokens

Input: $2.20/M

Output: $6.60/M

Try in Chat Get API Key

Specifications

Modalities

Input

text

image

file

Output

text

Supported Parameters

include_reasoning

logprobs

max_tokens

reasoning

reasoning_effort

response_format

seed

structured_outputs

temperature

tool_choice

tools

top_logprobs

top_p

Reasoning Configuration

Default: Thinking on (effort: high)
Selectable effort levels: xhigh
high
medium
low
Turning thinking off: Not possible (always thinks)

How to control thinking, and what it costs

Data policy

Prompt retention: May be retained
Training: Not used for training

Whether a provider trains on prompts is a declared value from our terms with them. Retention is determined from the upstream listing for every host this model can reach. Neither is guessed.

Measured performance

Time to first response (p50)1.3s

Throughput (p50)88 tokens/s

Measured through FastMetal's own gateway over the last 30 days (654 first-response samples).

Throughput is sampled once a day: the median of 26 observations over the last 30 days.

Time to first response is measured to the first streamed chunk; throughput counts all output tokens, including reasoning tokens. These are observations, not a performance guarantee.

Code Examples

curl https://api.fastmetal.ai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "grok-4.6",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

How Grok 4.6 actually answers

Real responses to our standard prompts, recorded on FastMetal.

Count the number of 'r's in 'strawberry'

Count the number of 'r's in 'strawberry'. Explain your reasoning step by step.

The word "strawberry" consists of the following letters in order: s-t-r-a-w-b-e-r-r-y Checking each one: - s: not r - t: not r - r: yes (count = 1) - a: not r - w: not r - b: not r - e: not r - r: yes (count = 2) - r: yes (count = 3) - y: not r There are exactly 3 r's. Final answer: 3

3.9s response451 output tokens

Debug This Error

I'm getting the following error in my Node.js application: TypeError: Cannot read properties of undefined (reading 'map') at UserList (/app/components/UserList.js:12:25) at renderWithHooks (/app/node_modules/rea…

**The error happens because `users` is `undefined` when `.map()` is called.** `UserList` destructures `{ users }` from its props. If the parent doesn't pass a `users` prop (or passes `undefined`/`null`), `users` is `undefined` and `undefined.map` throws. This is extremely common when: - Data is loaded asynchronously (initial render happens before the fetch finishes). - The parent component forgets to pass the prop or uses a different name. - The component is rendered before the data is ready (e.g. in a loading state). ### Quick fixes **1. Default to an empty array (recommended)** ```jsx function UserList({ users = [] }) { return ( <ul> {users.map(user => ( <li key={us…

20.3s response598 output tokens

Code Review

Please review the following Python function and suggest improvements for readability, performance, and best practices: def get_data(url, retries=3): import requests import time for i in range(retries):…

Here is a focused review of the function, followed by a cleaned-up version. ## Issues **Imports inside the function** `requests` and `time` should live at module level. Importing on every call is slower, hides dependencies, and is harder to mock in tests. **Bare `except:`** This swallows `KeyboardInterrupt`, `SystemExit`, and programming errors. Catch `requests.RequestException` (network/timeout/HTTP adapter errors) and JSON/`KeyError`/`TypeError` separately. **Retrying the wrong things** A 404 or 400 will never succeed on retry. Retry transient failures (timeouts, connection errors, 429, 5xx). Fail fast on most 4xx responses. **Sleep after the last attempt** On the final failur…

27.9s response1770 output tokens

Compare these answers side by side with other models →

Frequently asked questions

How much does the Grok 4.6 API cost?: On FastMetal, Grok 4.6 is billed per token: $2.20 per 1M input tokens and $6.60 per 1M output tokens on a US-dollar account, before tax. Usage is drawn from a prepaid balance; there is no subscription or monthly fee.
Can I call Grok 4.6 with the OpenAI SDK?: Yes. Point base_url at https://api.fastmetal.ai/v1 and pass "grok-4.6" as the model; existing OpenAI-style code works unchanged, including streaming, tool calls and structured output.
What is the context window of Grok 4.6?: 500,000 tokens, shared between the prompt and the response.
What do I need to try Grok 4.6?: Create an account and add credit; the model is then available both in the browser chat and over the API. There is no contract or minimum spend.