Regional Endpoints

https://api.x.ai is the global endpoint. SpaceXAI may route requests between regions for capacity or reliability, so the processing location is not guaranteed.

If you need API request handling and model inference to happen in the United States, use the US regional endpoint, https://us.api.x.ai/v1. It is available to every team, and your existing API keys work on both endpoints.

Point your client at https://us.api.x.ai/v1; the Python SDK (xai_sdk) takes the bare host, us.api.x.ai.

import { createXai } from '@ai-sdk/xai';
import { generateText } from 'ai';

const xai = createXai({
  apiKey: process.env.XAI_API_KEY,
  baseURL: 'https://us.api.x.ai/v1',
});

const { text } = await generateText({
  model: xai.responses('grok-4.6'),
  prompt: 'Explain latency versus throughput in two sentences.',
});

console.log(text);

import os

from xai_sdk import Client
from xai_sdk.chat import user

client = Client(
    api_key=os.getenv("XAI_API_KEY"),
    api_host="us.api.x.ai",
)

chat = client.chat.create(model="grok-4.6")
chat.append(user("Explain latency versus throughput in two sentences."))

print(chat.sample().content)

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.getenv("XAI_API_KEY"),
    base_url="https://us.api.x.ai/v1",
)

response = client.responses.create(
    model="grok-4.6",
    input="Explain latency versus throughput in two sentences.",
)

print(response.output_text)

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.XAI_API_KEY,
  baseURL: 'https://us.api.x.ai/v1',
});

const response = await client.responses.create({
  model: 'grok-4.6',
  input: 'Explain latency versus throughput in two sentences.',
});

console.log(response.output_text);

curl https://us.api.x.ai/v1/responses \
  -H "Authorization: Bearer $XAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "grok-4.6",
    "input": "Explain latency versus throughput in two sentences."
  }'

grok-4.6 is currently the only model available on the US endpoint; the models page in the console and GET https://us.api.x.ai/v1/models always show the current list. Requesting a model that is not on that list, including grok-latest, fails with 404 Not Found:

{
  "code": "not-found",
  "error": "The model grok-4-1-fast-reasoning does not exist or your team <team_id> does not have access to it. If you believe this is a mistake, please contact support and quote your team ID and the model name."
}

If a request for a globally available model returns this error, confirm that your client is calling the intended endpoint before troubleshooting model access. The image generation, video generation, and voice APIs are not served by the US endpoint; use the global endpoint for those.

Token usage on the US endpoint costs 10% more than on the global endpoint. The premium applies to input, output, and cached input tokens, including long-context rates, and prompt caching discounts still apply. The current per-token rates are on each model's detail page, reached from the models page, and on the Pricing page.

When you call https://us.api.x.ai/v1, SpaceXAI guarantees that the following happen in the United States:

• Handling of the request by SpaceXAI's API servers.
• Inference for the model you request.
• Safety moderation of the request and the response.
• Storage of the request metadata, prompt inputs, and model outputs that SpaceXAI retains. The Security FAQ describes what is retained and for how long.

The image generation, video generation, and voice APIs are not served by the US endpoint. Files, Collections, and server-side tools such as web search, X search, and code execution still work on the US endpoint, but they are outside the US guarantee and may process data outside the United States. If your requirements cover these features as well, avoid them when calling the US endpoint, or contact support@x.ai to discuss your configuration.

The guarantee also does not cover the network path between your own users or infrastructure and the endpoint.

Yes. Requests to the US endpoint accept the same parameters as the global endpoint, including function calling, server-side tools, file attachments, and structured outputs. However, server-side tools and files are outside the US guarantee; see What the guarantee covers.

Prompt cache hits are not guaranteed across endpoints. Keep each conversation on one endpoint and set a prompt_cache_key so its requests are routed together.

Yes. Zero Data Retention is a team-level setting, so it applies to every request your team makes regardless of the endpoint.