# Grok not working today: fixes for every error xAI documents

> Check status.x.ai first, then match the error code to the fix in xAI's own docs, from a 401 key problem to a 429 rate limit.

URL: https://claudegpt.si/news/grok-not-working-today-fixes  
Published: 2026-11-02T11:00:56.743Z  
Author: ClaudeGPT Si Newsroom  
Section: Guides  
Topics: Grok, xAI, Troubleshooting, API  
Publisher: ClaudeGPT Si

## Key takeaways

- xAI tells users to check status.x.ai for outages before changing anything.
- A 429 means a rate limit was hit; xAI recommends exponential backoff and checking limits in the Console.
- A 401 means the Authorization Bearer header is missing or invalid, and a new key can be made in the xAI Console.

Grok not working today? Start with the xAI status page at status.x.ai, then match the error you see to the list below. The fixes here come from the xAI developer docs, which cover the Grok API. They are the only troubleshooting steps xAI publishes in the pages we could read, so this guide is for API and developer errors, not for the Grok chat app.

If you use Grok through the website or the X app, see our guide to [Grok's weekly limit](https://claudegpt.si/news/grok-weekly-limit) and whether [Grok is free](https://claudegpt.si/news/is-grok-free). Those cover the usage caps that most often stop the chat app from answering.

## Check whether Grok is down

xAI's [debugging page](https://docs.x.ai/developers/debugging) says to check [status.x.ai](https://status.x.ai) for ongoing outages. The status page also offers an RSS feed, so you can be told when a problem starts or ends. If an incident is listed there, the fix is to wait. Nothing in your code or account will change it.

## What each Grok API error means

When a request works, you get 200 OK. Otherwise the response carries one of these codes, which xAI documents with a cause and a fix (as of November 2, 2026).

| Code | What it means | What xAI says to do |
| --- | --- | --- |
| 400 Bad Request | Something is wrong with the request | Check the request body and the URL |
| 401 Unauthorized | The authorization header is missing or invalid | Send `Authorization: Bearer <XAI_API_KEY>`. Get a new key in the xAI Console if needed |
| 403 Forbidden | Your account lacks permission for this action | Ask your team admin |
| 404 Not Found | The endpoint or something in the body was not found | Check the body and URL against the API Reference |
| 405 Method Not Allowed | The endpoint does not accept that HTTP method | Check the method in the API Reference, for example POST sent to a GET-only endpoint |
| 415 Unsupported Media Type | Applies to POST endpoints | The page gives no cause or fix |
| 422 Unprocessable Entity | A field in the POST body has an invalid format | Check the body against the API Reference |
| 429 Too Many Requests | You hit a rate limit | Slow down, or raise your limit in the xAI Console |

One success code is easy to mistake for a failure. A 202 Accepted comes from the deferred chat completion endpoint and means the request is queued but not finished. Wait and check again.

## How to fix a 401 error

1. Confirm the request sends the header `Authorization: Bearer <XAI_API_KEY>`.
2. Check that the key is the one you meant to use and was copied in full.
3. If it still fails, create a new key in the xAI Console and replace the old one.

## How to fix a 429 error

A 429 means you exceeded a limit. According to the xAI [rate limits page](https://docs.x.ai/developers/rate-limits), each model has a cap on requests per second and on tokens per minute, and both depend on your team's tier. Every token counts toward the per-minute cap: prompt, completion, reasoning and cached tokens.

1. Open the Models page in the xAI Console. It shows your current tier and the limits for each model.
2. Retry with exponential backoff. The docs suggest waits of 1, 2, 4 and 8 seconds, and giving up after a set number of attempts.
3. If you need more room, raise your tier or request an increase in the Console. For large capacity, the docs point to sales@x.ai.

Tiers are based on cumulative API spend since January 1, 2026, run from Tier 0 ($0) to Tier 4 ($5,000), and upgrade automatically. A tier never goes down once reached. With the xAI SDK, the same problem shows up as the gRPC error `RESOURCE_EXHAUSTED`.

## How to fix 400, 404 and 422 errors

These are problems with the request, so retrying will not help. Compare your endpoint, method and JSON body with the API Reference, field by field. A 422 in particular points to one field with an invalid format.

## If nothing here fixes it

xAI says to email support@x.ai with your request, the response and the relevant logs, or to ask in the #help channel of the xAI API Developer Discord.

## What this guide does not cover

xAI's help pages for the consumer Grok app on X were not readable when we checked on November 2, 2026, so no app-specific fixes are included here. We did not add steps from third-party sites.

## Frequently asked questions

### Is Grok down right now?

Check status.x.ai, which xAI names as the place for ongoing outages. It also has an RSS feed. If no incident is listed, the problem is more likely in your request or account.

### What does a 429 error mean in Grok?

It means you exceeded a rate limit for requests per second or tokens per minute. xAI recommends retrying with exponential backoff and checking your limits on the Models page in the xAI Console.

### Why does the Grok API return 401 Unauthorized?

The authorization header is missing or invalid. Send Authorization: Bearer with your xAI API key, and create a new key in the xAI Console if needed.

### Who do I contact if Grok still does not work?

xAI says to email support@x.ai with your request, response and logs, or ask in the #help channel of the xAI API Developer Discord.

## Sources

- [Debugging errors](https://docs.x.ai/developers/debugging), xAI
- [Rate limits](https://docs.x.ai/developers/rate-limits), xAI
- [xAI status](https://status.x.ai), xAI