• Sources: primary
  • Summary: GitHub began investigating degraded Copilot AI Model Providers at 17:46 UTC, attributed the GPT-5.4 Nano errors to an upstream model provider at 17:58, reported mitigation at 18:10 and resolved the incident at 18:24. GitHub says a root cause analysis will follow.
  • Why it matters: Copilot users relying on GPT-5.4 Nano saw failures, which shows availability depends on the upstream provider of each model.