Skip to main content
Baseten logo

Is Baseten Down?

No — Baseten is up

Reachable from all 8 checked regions

Average response time: 412ms

Last checked · checks run every 6 hours

Official status page: https://status.baseten.co

Baseten uptime

100%
Last 7 days
100%
Last 30 days
100%
Last 90 days
369ms
Avg response, 30 days

Measured from multiple regions every 6 hours. Percentages count only checks that returned an availability answer — 8 days measured so far. A dash means that window does not yet hold enough measured days to publish a figure.

30-day history

8-day clean streak
Sep 7: no data
Sep 8: no data
Sep 9: no data
Sep 10: no data
Sep 11: no data
Sep 12: no data
Sep 13: no data
Sep 14: no data
Sep 15: no data
Sep 16: no data
Sep 17: no data
Sep 18: no data
Sep 19: no data
Sep 20: no data
Sep 21: no data
Sep 22: no data
Sep 23: no data
Sep 24: no data
Sep 25: no data
Sep 26: no data
Sep 27: no data
Sep 28: no data
Sep 29: 100.00% uptime, 23 checks
Sep 30: 100.00% uptime, 28 checks
Oct 1: 100.00% uptime, 30 checks
Oct 2: 100.00% uptime, 32 checks
Oct 3: 100.00% uptime, 32 checks
Oct 4: 100.00% uptime, 32 checks
Oct 5: 100.00% uptime, 32 checks
Oct 6: 100.00% uptime, 24 checks
Sep 7 Today
No downtime Partial Downtime Not measurable No data

Reachability by region

Each region runs its own request from a different part of the world. A service can be up for one continent and down for another, which is usually the first sign of a routing or CDN problem.

ams
214ms
DNS 59ms TCP 8ms TLS 26ms TTFB 187ms
arn
249ms
DNS 113ms TCP 1ms TLS 11ms TTFB 240ms
cdg
585ms
DNS 463ms TCP 8ms TLS 15ms TTFB 555ms
dfw
331ms
DNS 152ms TCP 2ms TLS 65ms TTFB 282ms
ewr
259ms
DNS 131ms TCP 1ms TLS 20ms TTFB 246ms
sjc
340ms
DNS 277ms TCP 1ms TLS 8ms TTFB 333ms
syd
989ms
DNS 352ms TCP 1ms TLS 17ms TTFB 981ms
yyz
329ms
DNS 87ms TCP 12ms TLS 35ms TTFB 279ms

What Baseten does

Baseten is an inference platform that AI product teams use to serve open-source and custom models in production. Companies deploy their own models to dedicated GPUs, packaged with its Truss framework, or call hosted open models through Model APIs, an OpenAI-compatible endpoint at inference.baseten.co. It also offers training and async inference, so an outage can take down the AI features of apps built on top of it.

What an outage looks like

Most incidents are narrow. Requests to models in one cluster or region return bursts of HTTP 500 or 502 errors while other deployments answer normally, or a single hosted model on Model APIs fails while dedicated deployments keep working. Some incidents only affect the web app and metrics: charts in the UI and API stop updating while inference keeps serving. Scale-ups can also run slower than usual.

What to do about it

Check status.baseten.co, which splits Dedicated Inference, Model APIs, Training, Model Management API, Web Application and Homepage and Docs. Async requests are retried automatically on 500, 502, 503 and 504 errors, up to 3 attempts by default. Webhook results get only one retry and are lost if both attempts fail, so write outputs to storage in your model's postprocess function.

Is it down for everyone, or just you?

If this page says Baseten 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

Together AI
Together AI
F
fal
Replicate
Replicate
Groq
Groq

Baseten outage FAQ

Is Baseten down for everyone or just me?
Check which component status.baseten.co marks, because most Baseten incidents hit one cluster, one region or one hosted model. If the page is green, open your deployment in the Baseten dashboard: an Unhealthy status means the deployment is active but hitting runtime errors, and Baseten's docs say to check the model logs, then try deactivating and reactivating it.
Does a Baseten outage affect both Model APIs and dedicated deployments?
Not usually. Baseten lists Model APIs and Dedicated Inference as separate components, and many incident titles say one is not affected, such as a partial Nemotron Ultra Model API outage marked Dedicated Inference is not affected. Metrics incidents in the UI and API are often labelled Inference not affected, so a blank dashboard chart does not mean your model is down.
Will Baseten retry my async requests during an outage?
Yes, for server errors. Baseten retries failed async inference on 500, 502, 503 and 504 responses with exponential backoff, 3 attempts by default and up to 10. Errors such as 422 or 404 fail immediately. Queued requests wait until they run or expire, after 10 minutes by default and up to 72 hours with max_time_in_queue_seconds.
Can I lose async results if my webhook is down?
Yes. Baseten makes two webhook delivery attempts, one initial and one retry about 2 seconds later, each with a 10-second timeout, and its docs say the output is permanently lost if both fail. The status endpoint shows WEBHOOK_FAILED but never returns the output, so save results to cloud storage in your model's postprocess function.
How long do Baseten outages usually last?
Most posted incidents are resolved within an hour. In September 2026 a burst of 500s for models in GCP central1 lasted about 7 minutes, one US-East cluster about 45 minutes, and a Kimi K3 Model API outage just under an hour. A spike in 500s in a single us-central region on September 29 ran nearly 12 hours. Timelines are on status.baseten.co.

How we measure this

  • We request Baseten's public endpoint every 6 hours from Fly.io regions across six continents — the most recent check ran from 8 of them.
  • 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. We only change the 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 Baseten publishes an official status feed we read it too. An all-clear from the vendor can soften an unconfirmed degradation; a vendor-declared outage only worsens our verdict when our own checks corroborate it.

Get alerted when Baseten goes down

This page refreshes every 6 hours. 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