Live status for every AI provider

AI API status

Live status, latency and uptime for AI model APIs, inference platforms, voice and image APIs and GPU clouds.

OpenAI APIClaude APIGemini APIGroqOpenRouterElevenLabs

LiveTiming every AI API, around the clock_

AI status board_

Every AI provider, live

API response time against its own 24 hour average, the last 24 hours as a chart, and uptime from our checks every few minutes. Official status where the provider publishes it.

Sort

AI model APIs

Inference platforms and routers

Voice and speech APIs

Image and video APIs

Search, embeddings and data APIs

GPU clouds

AI API outage questions

Is the AI API down right now?

This page checks the major AI providers live: their official status pages where they publish one, plus our own timing of each API every few minutes. The board above puts any provider with a problem first.

Why is my AI API call failing?

Most failures are not outages. A 401 or 403 means the API key is wrong or out of credit, a 429 means a rate limit or quota, and a 404 often means a retired model name. 500, 502, 503 and 529 errors are on the provider's side: retry with exponential backoff and check the status here.

What should I do when my AI provider is down?

Retry with backoff for short incidents. For longer ones, switch to another provider that serves a similar model (each provider page lists alternatives), or route requests through a gateway that falls back automatically.

How do you measure API latency?

Every few minutes our servers send a request to each provider's API host and time the answer. Without an API key that answer is usually a quick "unauthorized", so the time shows how fast the provider's API edge responds, not how long a model takes to generate a reply.