Baseten Outage History

73 incidents recorded over the last 365 days, 70 of them resolved. Durations are measured from when PulsAPI first saw the problem to when it cleared, which is usually longer than the vendor's own figure.

Live Baseten status98.87% uptime history

Every Recorded Baseten Incident

Baseten incidents, newest first, with date, duration and status.
DateIncidentDurationStatus
Aug 19, 2026Elevated errors for moonshotai/Kimi-K3 for US-pinned traffic on Model APIs. **Dedicated Inference is not affected**This incident has been resolved.36 minResolved
Aug 18, 2026Elevated errors for moonshotai/Kimi-K3 for US-pinned traffic on Model APIs. **Dedicated Inference is not affected**This incident has been resolved.1h 17mResolved
Aug 18, 2026Elevated errors for moonshotai/Kimi-K3 for US-pinned traffic on Model APIs. **Dedicated Inference is not affected**This incident has been resolved.11 minResolved
Aug 18, 2026Elevated errors for moonshotai/Kimi-K3 for US-pinned traffic on Model APIs. **Dedicated Inference is not affected**This incident has been resolved.47 minResolved
Aug 17, 2026Elevated errors for moonshotai/Kimi-K3 for US-pinned trafficThis incident has been resolved.45 minResolved
Aug 17, 2026Metrics degradation in UI and API. **Inference not affected**This incident has been resolved.24 minResolved
Aug 6, 2026Database Upgrade. Inference will not be affectedThe scheduled maintenance has been completed.1h 0mScheduled
Aug 4, 2026Database Upgrade. Inference will not be affectedThe scheduled maintenance has been completed.1h 0mScheduled
Aug 1, 2026Elevated inference 502s for a subset of models in a single US clusterThis incident has been resolved.43 minResolved
Jul 31, 2026Elevated inference 502s for a subset of models in a single US clusterThis incident has been resolved.28 minResolved
Jul 31, 2026Slower than usual scale-ups for B200sThis incident has been resolved.11 minResolved
Jul 22, 2026Elevated inference failures for models in an EU clusterThis incident has been resolved.32 minResolved
Jul 22, 2026Slower than usual scale-ups in 1 us-central clusterThis incident has been resolved.19 minResolved
Jul 18, 2026GLM 5.2 Model API partial outage due to provider network failureThis incident has been resolved.32 minResolved
Jul 18, 2026Elevated inference errors in 1 Canada cluster due to provider network outageThis incident has been resolved.5 minResolved
Jul 16, 2026Elevated errors in accessing the Baseten UI. **Inference is not affected**This incident has been resolved.1h 18mResolved
Jul 13, 2026Partial Outage on Model APIsKimi-K2.6 recovered.2h 6mResolved
Jul 9, 2026Elevated inference 500s for models deployed in 1 US-east clusterThis incident has been resolved.20 minResolved
Jul 8, 2026Metrics degradation in UI and API. **Inference not affected**This incident has been resolved.1h 55mResolved
Jul 6, 2026Elevated inference 500s — network outage in our cloud provider in DelawareThis incident has been resolved.13 minResolved
Jul 3, 2026Elevated errors for GLM 5.2 on Model APIs. **Dedicated Inference is not affected**This incident has been resolved.1h 25mResolved
Jul 3, 2026Elevated inference 500s for a subset of models deployed in 1 clusterThis incident has been resolved.15 minResolved
Jul 2, 2026Degraded metrics performance in dashboard and API. **Inference is not affected**This incident has been resolved.1h 13mResolved
Jul 1, 2026Elevated errors in a single EU clusterThis incident has been resolved.59 minResolved
Jun 29, 2026Elevated error rates on GLM 5.2 Model APIs. **Dedicated Inference is not affected**This incident has been resolved.35 minResolved
Jun 27, 2026Elevated inference 500s for models in 1 cluster due to cloud provider outageThis incident has been resolved.15 minResolved
Jun 26, 2026Elevated model deployment failures. **Inference is not affected**The incident has resolved. New model deployments are successfully coming up.1h 14mResolved
Jun 26, 2026Elevated errors for dedicated inference in a subset of clustersThis incident has been resolved.14 minResolved
Jun 23, 2026Slower than usual scale-ups. High failure rate for new model deploys. Existing replicas aren't impacted.This incident has been resolved.44 minResolved
Jun 18, 2026Elevated errors for models in 1 Austin-based cluster affecting some models on B200sThis incident has been resolved.1h 43mResolved
Jun 17, 2026Elevated errors for models in 1 Sweden-based clusterThis incident has been resolved.1h 23mResolved
Jun 15, 2026Elevated errors for models in a single EU cloud providerThis incident has been resolved.44 minResolved
Jun 13, 2026Elevated errors for models in a single cloud providerThis incident has been resolved.44 minResolved
Jun 5, 2026Elevated errors in a single US-East clusterThis incident has been resolved.19 minResolved
Jun 4, 2026Elevated latencies for a subset of models in a US-East clusterThis incident has been resolved.1h 6mResolved
Jun 3, 2026Slower than usual scale-ups for some models using H100sThis incident has been resolved.2h 14mResolved
May 29, 2026Elevated async inference errors in a single clusterThis incident has been resolved.32 minResolved
May 26, 2026Slower than usual scale-ups in 1 India clusterThis incident has been resolved.13 minResolved
May 20, 2026Elevated errors for models in single regionThis incident has been resolved.13 minResolved
May 18, 2026Slower than usual scale-ups for L4s in US-EastThis incident has been resolved.38 minResolved
May 16, 2026Slower than usual scale-ups in EU due to a provider failureThis incident has been resolved.1h 32mResolved
May 16, 2026Higher latencies for a subset of models in one cluster. Root caused and a fix is in progressThe incident is resolved.36 minResolved
May 15, 2026Elevated errors in a single clusterThis incident has been resolved.2h 0mResolved
May 10, 2026Training cluster is down due to a provider failureThe issue has been resolved.15 minResolved
May 8, 2026Intermittent failure in loading the webapp — inference is *not* affectedThe issue is resolved.41 minResolved
May 7, 2026Elevated errors for models in single regionThis incident has been resolved.47 minResolved
Apr 30, 2026Planned Maintenance on our Training Cluster. Inference will not be affected.The scheduled maintenance has been completed.3h 0mScheduled
Apr 13, 2026Slower than usual scale-ups for a small subset of modelsThis incident has been resolved.49 minResolved
Apr 4, 2026Slower than usual scale-ups for a small subset of modelsThis incident has been resolved.4h 45mResolved
Apr 3, 2026Builds of new models are currently failing. **Inference is not affected**This incident has been resolved.55 minResolved
Apr 2, 2026Elevated deploy failures. Inference is **not** affected.There was a backup in the deploy queue, resulting in some deploy failures. Please retry any deploys from this period.-67 minResolved
Mar 12, 2026gpt-oss model API service degradationFix is in place and the issue has been resolved.17 minResolved
Feb 28, 2026Elevated 500s on Model APIs. **Dedicated Inference is not affected**This incident has been resolved.42 minResolved
Feb 27, 2026Partial inference outage in AWS us-east-1This incident has been resolved.54 minResolved
Feb 13, 2026Network Errors in single EU ClusterThis incident has been resolved.3h 25mResolved
Feb 9, 2026Elevated inference latencies in AWS us-west-2This incident has been resolved.2h 13mResolved
Jan 22, 2026Baseten web app is down. **Inference is unaffected**This incident has been resolved.23 minResolved
Jan 21, 2026Elevated latency for inference requests in single region in Baseten Cloud WestThis incident has been resolved.1h 3mResolved
Dec 21, 2025GPT-OSS on Model APIs is degraded. **Dedicated inference is not affected**.This incident has been resolved.21 minResolved
Dec 21, 2025Slower than usual scale-ups and new deploys for all models in GCPThis incident has been resolved.54 minResolved
Dec 6, 2025Deepseek V3-0324 on Model APIs is down. **Dedicated inference is not affected**.The incident is resolved.20 minResolved
Nov 17, 2025Elevated error rates for inference requests for a subset of models in GCP us-central1This incident has been resolved.19 minResolved
Nov 3, 2025New builds are failing to push. Inference is *not* affectedThis incident has been resolved.15 minResolved
Nov 1, 2025Elevated errors and delays in new model deploys in 1 cluster. Inference is *not* affected.This incident has been resolved.38 minResolved
Oct 30, 2025Elevated errors in model deploys and promotions in 2 clusters. Inference is *not* affected.This incident has been resolved.8 minResolved
Oct 25, 2025Slower than usual scale-ups in 1 regionThis incident has been resolved.3h 42mResolved
Oct 24, 2025Slower than usual scale-ups in 2 regionsThis incident has been resolved.4h 13mResolved
Oct 23, 2025Model logs are slow to load in a single cluster. Inference is not affected.This incident has been resolved.15 minResolved
Oct 20, 2025Login timing out on webappThe issue with our auth provider has been fixed, and logins work again.6h 37mResolved
Oct 20, 2025Baseten Web Application is down due to AWS outage. Inference is unaffected.Upstream provider has recovered, and the Baseten web app is available.2h 15mResolved
Oct 13, 2025High latency on Baseten Web Application and Model Management APIs. Inference is unaffected.The issue is resolved.14 minResolved
Sep 25, 2025Build Infrastructure degraded due to Dockerhub outageThis incident has been resolved.57 minResolved
Sep 22, 2025L4 capacity in single us-east region having issuesThis incident has been resolved.12 minResolved

How to Read This Baseten Incident Log

Each row is an incident PulsAPI observed, not a summary written afterwards. The duration is wall-clock time between the first failing check and the first clean one, so it includes the window before Baseten acknowledged anything. Vendor post-mortems typically measure from acknowledgement, which is why their numbers are usually shorter.

Incidents still open have no duration yet and are listed as ongoing rather than being given a running total. The archive covers the last 365 days; anything older has aged out of the window rather than never having happened.

The longest single Baseten outage in this window ran 6h 37m, against a mean recovery of 1h 4m. If you depend on Baseten in a customer-facing path, the longest figure is the one to design around. The mean is what happens on a normal bad day; the maximum is what happens on the worst one.