# Is OpenAI API Down? Status and Error Checks

TL;DR: Check whether an OpenAI API problem looks like an outage, an invalid key, a forbidden model, or a rate limit issue.

Official status: https://status.openai.com/
Public endpoint: https://api.openai.com/v1/models
Run diagnosis: https://llmping.app/#diagnose

## Honest boundary

llmping does not ping user API keys and does not claim fake real-time provider status. Use the checks below to separate outage symptoms from key, quota, permission, and runtime problems.

## Historical uptime

llmping samples provider public endpoints and official status feeds every 5 minutes with a Cloudflare Worker and stores real samples in D1. The status page displays 7, 30, and 90 day public endpoint uptime once samples exist.

Uptime API: https://llmping-uptime-worker.stcatzlee.workers.dev/api/uptime?provider=openai

## Quick checks

- 401 means your key is missing or invalid; it does not prove OpenAI is down.
- 403 means your key is recognized but forbidden for the requested action or model.
- 429 means rate limit or quota pressure; check whether the body says quota or RPM/TPM.
- 5xx across multiple keys and regions is stronger outage evidence.

## cURL test command

```bash
curl -i https://api.openai.com/v1/models \
  -H "Authorization: Bearer $OPENAI_API_KEY"
```

## Related pages

- https://llmping.app/errors/openai/429/
- https://llmping.app/errors/openai/401/
- https://llmping.app/errors/openai/403/