Is Cerebras Down?

No — Cerebras is up

Cerebras's status page: Operational · no open incidents
FlareWarden checks: Website reachable · 8 of 8 regions · 853 ms

Right now

The verdict is the worse of these two. How we decide

Cerebras's status page

status.cerebras.ai · read 4 minutes ago · checked every 3 minutes

Operational

Cerebras's components, as its status page reports them
Component	Status
GPT-OSS-120B	Operational
Developer Console	Operational
Qwen-3.8-27b	Operational

Open their status page

FlareWarden checks

cerebras.ai from 8 regions · when you open this page (at most every 30 minutes), otherwise every 6 hours

Reachable

Regions up: 8 of 8
Response: 853 ms
Checked: 2 hours ago

By region

Amsterdam reachable, 795 ms
Chicago reachable, 478 ms
San Jose reachable, 711 ms
Singapore reachable, 1317 ms
Stockholm reachable, 822 ms
Sydney reachable, 1258 ms
Tokyo reachable, 917 ms
Toronto reachable, 531 ms

Timing by region

Connection timing for each region that loaded the page
Region	DNS	TCP	TLS	TTFB
Amsterdam	163 ms	6 ms	18 ms	516 ms
Chicago	181 ms	11 ms	15 ms	430 ms
San Jose	139 ms	2 ms	9 ms	464 ms
Singapore	219 ms	2 ms	9 ms	657 ms
Stockholm	115 ms	1 ms	10 ms	476 ms
Sydney	207 ms	1 ms	5 ms	642 ms
Tokyo	122 ms	2 ms	8 ms	467 ms
Toronto	205 ms	11 ms	29 ms	482 ms

We check that cerebras.ai loads, not every feature behind it. Test it yourself

History

Incidents Cerebras reported

Since Oct 22, 15 days · from status.cerebras.ai

Incidents: 1
Major or critical: 0
Median to resolve: 2h 33m

Partial Service Disruption
Unrated · Nov 3
lasted 2h 33m

Cerebras website availability

Measured by FlareWarden · cerebras.ai

Last 7 days: 100%
Last 30 days: 100%
Last 90 days: 100%
Avg response, 30 days: 818ms

Last 30 days

21-day clean streak

Oct 7 Today

No downtime Partial Downtime Not measurable No data

Measured by FlareWarden's own checks of cerebras.ai from multiple regions every 6 hours. Cerebras's status page sets the verdict above when it reports trouble, but never these figures. Percentages count only checks that returned an availability answer — 21 days measured so far. A dash means that window does not yet hold enough measured days to publish a figure.

What Cerebras does

Cerebras Inference is the cloud API that Cerebras Systems runs on its own wafer-scale chips, sold to developers who need very fast token generation. Applications call open models such as gpt-oss-120b and qwen-3.8-27b at api.cerebras.ai through an OpenAI-compatible interface, test prompts in the Cloud Console Playground, or reserve capacity for their organization with Dedicated Inference.

What an outage looks like

Requests to api.cerebras.ai return 500 InternalServerError or 503 ServiceUnavailable, or the SDK raises APIConnectionError when it cannot reach the API at all. Calls that hang end at the SDK's default one-minute timeout. The status page tracks each model and the Developer Console separately, so one model can fail while another answers. A 429 says whether your uncached or total token limit ran out.

What to do about it

Check status.cerebras.ai, which lists GPT-OSS-120B, Qwen-3.8-27b and the Developer Console as separate rows. The Cerebras SDK already retries connection errors, 408, 429 and 5xx responses twice with exponential backoff, so raise max_retries rather than wrapping it in a second retry loop. If one model is degraded, switch to the other Shared Inference model. Report persistent failures to support@cerebras.net.

Is it down for everyone, or just you?

If this page says Cerebras is up but it is not loading for you, the problem is between you and them. Run a check against any URL from up to 17 cities across three regions to find out where it breaks.

Test it yourself

Related services

Cerebras outage FAQ

Is Cerebras down for everyone or just me?

Check the row for your model on status.cerebras.ai first, because Cerebras lists each Shared Inference model and the Developer Console separately. If that row is green, read the error: a 429 is your organization's rate limit and a 401 is an API key problem. The checks on this page probe www.cerebras.ai, the company website, so an API fault on api.cerebras.ai can happen while they stay green.

Why do I get 429 errors from Cerebras after buying credits?

Because Developer tier limits still apply per model, and other limits can bite first. Cerebras says an Org Admin may have set a lower project-level limit, that a large max_completion_tokens value counts against your quota before the request runs, and that only uncached tokens count toward the uncached limit. The 429 message names which limit was exceeded, and the Limits page in the Cloud Console shows your actual quotas.

Does the Cerebras SDK retry failed requests on its own?

Yes, twice by default with a short exponential backoff. Cerebras documents that its SDK retries connection errors, 408 Request Timeout, 429 rate limits and any 500-range error, and that timed-out requests are retried too. You can change this with the max_retries option, per client or per request. Requests time out after one minute by default, which the timeout option also changes.

Can I switch to another provider during a Cerebras outage?

Yes, if your code uses the OpenAI client library. Cerebras documents connecting it by setting base_url to https://api.cerebras.ai/v1 with a Cerebras API key, so moving to another OpenAI-compatible provider means changing the base URL, the key and the model ID. Cerebras notes that on gpt-oss-120b the same prompt may behave differently than on OpenAI, so test the fallback model before you need it.

Does the Cerebras status page cover Dedicated Inference?

Not directly. Its rows are the Shared Inference models and the Developer Console, and no row names a Dedicated Inference endpoint. For dedicated capacity, Cerebras offers Prometheus-compatible metrics for requests, tokens, latency and endpoint health, which show how your own endpoint is behaving. For reserved capacity or Enterprise limits, Cerebras points customers to its contact page or their account representative.

How we measure this

We request Cerebras's public endpoint every 6 hours from Fly.io regions across six continents — the most recent check ran from 8 of them. When you open this page and our last check is more than 30 minutes old, we check again on the spot; those checks update the verdict, never the availability figures.
A region counts as down only when it gets no usable HTTP response. A 403 or 429 means the origin answered and refused us, which we report as blocked, never as an outage.
A single failing region is treated as probe noise. Our checks only change their verdict when two consecutive cycles agree.
Response times average only the regions that actually served the page, so a timeout never inflates the number.
Where Cerebras publishes an official status feed we read it too, and the verdict is the worse of what it reports and what our checks see: a major outage of Cerebras itself is Down, trouble with one part of it is Degraded, problems at a few of a provider's many locations change nothing, and a security notice that marks nothing down is shown but changes nothing either. We check the feed every 3 minutes and read it again within minutes of any change to its incidents or components; a reading counts for 45 minutes after the last check that found nothing changed. The availability figures above are our measurement alone.

Get alerted when Cerebras goes down

We check Cerebras every 6 hours, and again when you open this page if that was more than 30 minutes ago. Your own monitors run as often as every 30 seconds, from up to 17 cities across three regions, and tell you the moment something breaks.

Start Free Monitoring

Free plan available No credit card required