Critical incident

GitHub: Disruption with some Copilot services

Affected: Copilot

Current status of GitHub: Operational — live GitHub status

Timeline

1. Oct 28, 2026, 14:12 UTC
   investigating
   We are investigating reports of impacted performance for some GitHub services.
2. Oct 28, 2026, 14:15 UTC
   investigating
   We are experiencing issues with upstream model providers resulting in a degraded experience across Copilot services. The team is actively investigating the issue.
3. Oct 28, 2026, 14:33 UTC
   investigating
   We are experiencing degraded availability for the GPT-5.3-Codex, GPT-5.6 Luna, GPT-6.1 Sol, GPT-5.6 Terra, GPT-6 Luna, and MAI Code 1.1 Flash models in Copilot products and IDE surfaces. This is due to an issue with an upstream model provider. While we work with them to resolve the issue we recommend choosing another model to continue using Copilot.
4. Oct 28, 2026, 14:55 UTC
   investigating
   We are experiencing degraded availability for the GPT-5.3-Codex, GPT-5.6 Luna, GPT-6.1 Sol, GPT-5.6 Terra, GPT-6 Luna, and MAI Code 1.1 Flash models in Copilot products and IDE surfaces. We have identified the root cause and working to apply a mitigation. While we work to resolve the issue we recommend choosing another model to continue using Copilot.
5. Oct 28, 2026, 15:15 UTC
   investigating
   We have applied the mitigation and are seeing signs of recovery. We will continue to monitor and expect full resolution soon.
6. Oct 28, 2026, 15:24 UTC
   investigating
   Copilot is experiencing degraded performance. We are continuing to investigate.
7. Oct 28, 2026, 15:26 UTC
   monitoring
   The degradation affecting Copilot has been mitigated. We are monitoring to ensure stability.
8. Oct 28, 2026, 15:28 UTC
   resolved
   On October 28, 2026, between 13:52 and 15:24 UTC, Copilot users saw errors when using several GPT-family models and the MAI Code 1.1 Flash model. Affected experiences included Copilot Chat in IDEs and on GitHub.com, the Copilot CLI, Copilot Code Review, the Copilot Coding Agent, Code Scanning Autofix, and Code Quality. About 9.73% of requests to the affected models failed, with a peak of about 36.97% at the start of the incident. The errors happened because a credential that we use to authenticate with an upstream model provider became invalid. We mitigated the incident by replacing the credential and redeploying the service. We are strengthening how we manage credentials for upstream providers to prevent this type of failure, returning retryable errors when an upstream provider rejects a request, and improving our monitoring so that we detect and mitigate this type of problem faster in the future.

Updates as published by GitHub (original).