Baseten outage history

Every smaller Baseten incident

Minor and short incidents from the Baseten status page, newest first, each with a line from the official updates. Times are UTC.

October 2026

• Oct 21, 2026, 01:48 UTC · Elevated errors on DeepSeek V4 Flash on Model APIs

  40 minutes · rated "Minor"

  "This incident has been resolved."

• Oct 21, 2026, 01:05 UTC · Elevated errors for image requests on multiple Model APIs

  21 minutes · rated "Minor"

  "There are elevated errors for image requests on Deepseek V4 Flash and GLM 5.3 Flash"

• Oct 20, 2026, 19:04 UTC · DeepSeek V4.1 Flash Fast Outage on Model APIs

  6 minutes · rated "Major"

  "This incident has been resolved."

• Oct 20, 2026, 01:35 UTC · Elevated errors on Model APIs

  8 minutes · rated "Minor"

  "Elevated errors across Model APIs due to provider outage"

• Oct 9, 2026, 21:29 UTC · Elevated errors on new model builds. Existing models are not affected

  37 minutes · rated "Minor"

  "A fix has been implemented and we are monitoring the results."

• Oct 6, 2026, 04:35 UTC · DeepSeek v4.1 outage on Model APIs. Dedicated Inference is not affected

  13 minutes · rated "Major"

  "This incident has been resolved."

September 2026

• Sep 29, 2026, 19:44 UTC · Elevated 500s in Single us-central Region

  13 minutes · rated "Major"

  "The issue has been identified and we're looking into a fix."

• Sep 29, 2026, 01:07 UTC · Metrics degradation in UI and API. Inference not affected

  21 minutes · rated "Minor"

  "We are continuing to investigate this issue."

• Sep 24, 2026, 18:35 UTC · Intermittent 5xx for a subset of models in 1 US Central cluster

  1 hour 22 min · rated "Minor"

  "0.15% of requests on this cluster resulted in a 5xx error, confined to two 2-3 min spikes. The team is continuing to investigate."

• Sep 24, 2026, 07:23 UTC · Elevated inference 500s for a subset of models deployed in 1 US-East cluster

  42 minutes · rated "Major"

  "The issue has been identified and a fix is being implemented."

• Sep 21, 2026, 18:34 UTC · Elevated inference 500s for a subset of models deployed in GCP central1

  7 minutes · rated "Major"

  "The issue has been identified and a fix is being implemented."

• Sep 21, 2026, 13:31 UTC · Metrics degradation in UI and API. Inference not affected

  11 minutes · rated "Minor"

  "This incident has been resolved."

• Sep 20, 2026, 07:19 UTC · Kimi K3 Model API outage

  55 minutes · rated "Major"

  "Calls to the US-hosted endpoint are resulting in errors. A fix is in progress."

• Sep 15, 2026, 01:22 UTC · Elevated errors for models in a single India-based cluster

  45 minutes · rated "Minor"

  "A fix has been implemented and we are monitoring the results."

• Sep 14, 2026, 16:21 UTC · Partial outage on Kimi K3 model API inference (non-US only)

  42 minutes · rated "Minor"

  "This incident has been resolved."

• Sep 9, 2026, 17:42 UTC · Partial outage on Nemotron Ultra Model API. Dedicated Inference is not affected

  6 minutes · rated "Minor"

  "A fix has been implemented and we are monitoring the results."

August 2026

• Aug 31, 2026, 06:38 UTC · Elevated inference 500s for models in a single US-based cluster

  24 minutes · rated "Major"

  "Fix is in place and traffic is back."

• Aug 27, 2026, 21:17 UTC · Elevated errors for models in a single US cluster due to provider outage

  14 minutes · rated "Minor"

  "Auto-recovery is in progress."

• Aug 25, 2026, 16:17 UTC · Elevated errors for deepseek-ai/DeepSeek-V4-Flash-0731 on Model APIs. Dedicated Inference is not affected

  41 minutes · rated "Minor"

  "A fix is in place and we are monitoring"

• Aug 25, 2026, 12:34 UTC · Elevated inference 500s for models in an EU cluster due to provider outage

  41 minutes · rated "Minor"

  "Failover in progress."

• Aug 19, 2026, 01:23 UTC · Elevated errors for moonshotai/Kimi-K3 for US-pinned traffic on Model APIs. Dedicated Inference is not affected

  37 minutes · rated "Minor"

  "A fix has been implemented and we are monitoring the results."

• Aug 18, 2026, 19:04 UTC · Elevated errors for moonshotai/Kimi-K3 for US-pinned traffic on Model APIs. Dedicated Inference is not affected

  1 hour 17 min · rated "Minor"

  "A fix has been implemented and we are monitoring the results."

• Aug 18, 2026, 17:24 UTC · Elevated errors for moonshotai/Kimi-K3 for US-pinned traffic on Model APIs. Dedicated Inference is not affected

  12 minutes · rated "Minor"

  "This incident has been resolved."

• Aug 18, 2026, 16:27 UTC · Elevated errors for moonshotai/Kimi-K3 for US-pinned traffic on Model APIs. Dedicated Inference is not affected

  48 minutes · rated "Minor"

  "This incident has been resolved."

• Aug 17, 2026, 16:56 UTC · Elevated errors for moonshotai/Kimi-K3 for US-pinned traffic

  46 minutes · rated "Minor"

  "This incident has been resolved."

• Aug 17, 2026, 13:11 UTC · Metrics degradation in UI and API. Inference not affected

  24 minutes · rated "Minor"

  "A fix has been implemented and we are monitoring the results."

• Aug 1, 2026, 23:57 UTC · Elevated inference 502s for a subset of models in a single US cluster

  43 minutes · rated "Minor"

  "The issue has been identified and a fix is being implemented."

July 2026

• Jul 31, 2026, 23:52 UTC · Elevated inference 502s for a subset of models in a single US cluster

  29 minutes · rated "Minor"

  "A fix has been implemented and we are monitoring the results."

• Jul 31, 2026, 16:46 UTC · Slower than usual scale-ups for B200s

  12 minutes · rated "Minor"

  "This incident has been resolved."

• Jul 22, 2026, 17:49 UTC · Elevated inference failures for models in an EU cluster

  33 minutes · rated "Minor"

  "We are continuing to work on a fix for this issue."

• Jul 22, 2026, 17:45 UTC · Slower than usual scale-ups in 1 us-central cluster

  19 minutes · rated "Minor"

  "The issue has been identified and a fix is being implemented."

• Jul 18, 2026, 22:40 UTC · GLM 5.2 Model API partial outage due to provider network failure

  32 minutes · rated "Major"

  "Escalated Dedicated inference is *not* affected."

• Jul 18, 2026, 00:22 UTC · Elevated inference errors in 1 Canada cluster due to provider network outage

  6 minutes · rated "Minor"

  "The issue has been identified and a fix is being implemented."

• Jul 16, 2026, 09:06 UTC · Elevated errors in accessing the Baseten UI. Inference is not affected

  1 hour 19 min · rated "Minor"

  "We are continuing to investigate this issue."

• Jul 13, 2026, 00:04 UTC · Partial Outage on Model APIs

  2 hours 7 min · rated "Major"

  "Kimi-K2.6 recovered."

• Jul 9, 2026, 14:14 UTC · Elevated inference 500s for models deployed in 1 US-east cluster

  14 minutes · rated "Minor"

  "The issue has been identified and a fix is being implemented."

• Jul 8, 2026, 16:50 UTC · Metrics degradation in UI and API. Inference not affected

  1 hour 55 min · rated "Minor"

  "Incorrect values recorded for some metrics"

• Jul 6, 2026, 21:08 UTC · Elevated inference 500s - network outage in our cloud provider in Delaware

  14 minutes · rated "Minor"

  "The issue has been identified and a fix is being implemented."

• Jul 3, 2026, 02:59 UTC · Elevated errors for GLM 5.2 on Model APIs. Dedicated Inference is not affected

  1 hour 26 min · rated "Major"

  "We are continuing to investigate this issue."

• Jul 3, 2026, 00:21 UTC · Elevated inference 500s for a subset of models deployed in 1 cluster

  16 minutes · rated "Minor"

  "We are continuing to work on a fix for this issue."

• Jul 2, 2026, 22:00 UTC · Degraded metrics performance in dashboard and API. Inference is not affected

  1 hour 14 min · rated "Minor"

  "A fix has been implemented and we are monitoring the results."

• Jul 1, 2026, 10:30 UTC · Elevated errors in a single EU cluster

  59 minutes · rated "Major"

  "Elevated errors due to degradation in a single EU cluster"

June 2026

• Jun 29, 2026, 18:26 UTC · Elevated error rates on GLM 5.2 Model APIs. Dedicated Inference is not affected

  35 minutes · rated "Minor"

  "Elevated error rates on GLM 5.2 Model APIs. Dedicated Inference is not affected"

• Jun 27, 2026, 06:35 UTC · Elevated inference 500s for models in 1 cluster due to cloud provider outage

  15 minutes · rated "Minor"

  "The issue has been identified and a fix is being implemented."

• Jun 26, 2026, 23:41 UTC · Elevated model deployment failures. Inference is not affected

  1 hour 14 min · rated "Minor"

  "The incident has resolved. New model deployments are successfully coming up."

• Jun 26, 2026, 22:00 UTC · Elevated errors for dedicated inference in a subset of clusters

  14 minutes · rated "Major"

  "We are continuing to work on a fix for this issue."

• Jun 23, 2026, 00:10 UTC · Slower than usual scale-ups. High failure rate for new model deploys. Existing replicas aren't impacted.

  44 minutes · rated "Minor"

  "We are continuing to work on a fix for this issue."

• Jun 18, 2026, 12:46 UTC · Elevated errors for models in 1 Austin-based cluster affecting some models on B200s

  1 hour 44 min · rated "Minor"

  "GLM 5.1 and Kimi k2.6 Model APIs are also affected."

• Jun 17, 2026, 04:40 UTC · Elevated errors for models in 1 Sweden-based cluster

  1 hour 23 min · rated "Minor"

  "The issue has been identified and a fix is being implemented."

• Jun 15, 2026, 16:14 UTC · Elevated errors for models in a single EU cloud provider

  44 minutes · rated "Minor"

  "A fix has been implemented and we are monitoring the results."

• Jun 13, 2026, 18:25 UTC · Elevated errors for models in a single cloud provider

  44 minutes · rated "Minor"

  "Network outage in single cloud provider (https://status.vultr.com/), elevated errors for some models"

• Jun 5, 2026, 02:10 UTC · Elevated errors in a single US-East cluster

  19 minutes · rated "Major"

  "This incident has been resolved."

• Jun 4, 2026, 20:53 UTC · Elevated latencies for a subset of models in a US-East cluster

  1 hour 6 min · rated "Minor"

  "The issue has been identified and a fix is being implemented."

• Jun 3, 2026, 14:56 UTC · Slower than usual scale-ups for some models using H100s

  2 hours 15 min · rated "Minor"

  "The issue has been identified and a fix is being implemented."