AI API status /

Is the Together AI API reachable right now?

A live probe of api.together.xyz from NetOkay’s edge, a test from your own browser, and why a model that has not been used recently can look like a network problem.

In short: The card below shows the latest probe of api.together.xyz from eight Cloudflare regions: one GET for /v1/models with a placeholder token and expects the API’s own 401. Then test from this browser: if the host answers on both sides but your first request to a model waits a long time, that is often a model loading, not your connection; a 30‑second idle limit on your path turns it into a failure.

Last updated

Where is api.together.xyz reachable from?

Checking…

Loading the latest round from every region…

Loading the latest observation of api.together.xyz…

From this browser

—

Not run. One request from your browser, sent only when you ask.

What this proves, and what it does not

Loading…

Together AI specifics

API host
api.together.xyz, HTTPS on port 443, OpenAI‑compatible under /v1/.
Streaming
Server-sent events with data: JSON chunks and data: [DONE]. Together also supports its own /inference shape, but OpenAI clients use the /v1/ surface unchanged.
Model start‑up
Serverless endpoints for less frequently used open models may need to load before the first token. That first‑byte wait can exceed idle limits on corporate proxies; the stream itself, once started, is fast.
Official status
status.together.ai lists incidents by service.
Client defaults
The Together Python SDK waits up to 600 seconds and retries up to 5 times with 0.5 to 8 second backoff, and records the CF-RAY response header as its request_id; the TypeScript SDK uses a 1-minute timeout and 2 retries. A model still loading can outlast the TypeScript default.

What Together AI status codes mean

Code Sent by Meaning
401 API application Invalid key. The API application produced this answer.
400 API application Unknown model name or invalid request body.
429 API application Rate limit reached.
503 API application A model is unavailable or still loading; the JSON body says so.
Every row is an HTTP answer: the host was reached and the request was refused or served. None of them means api.together.xyz is unreachable.

Reachable here but failing for you?

  1. Run the 30‑second idle stream from the network where your app fails. A model loading before its first token is a quiet connection, which is exactly what an idle limit cuts. Run the idle stream →
  2. Read the idle result. Closed during the silence points at an idle limit under 30 seconds on your path; the streaming guide has the fixes for proxies and gateways. Fixes in the streaming guide →
  3. Compare environments if this browser passes but a container, server or agent fails, then run the CLI from the failing environment against api.together.xyz. Compare environments →

Together AI API status FAQ

The first request to a model takes ages, then it is fast. Why?

The model was loading. This is a first‑byte delay rather than a streaming problem. If your client or proxy times out during that wait, raise the timeout or send a keep‑alive request first.

The probe is healthy but my dedicated endpoint fails. Does this page cover it?

No. Dedicated endpoints have their own hostnames. This page only observes the shared api.together.xyz host.

Does a 401 here mean Together is fine?

It means DNS, TLS and HTTP all work from Cloudflare’s edge to api.together.xyz and the API answered. It says nothing about model availability or your own network.