Incident with Copilot GPT-5.1-Codex
This summary is created by Generative AI and may differ from the actual content.
Overview
Degradation of the GPT 5.1 Codex model in Copilot service between 07:30 UTC and 11:21 UTC on February 20, 2026, resulting in a 4.5% error rate for users. The issue was isolated to the GPT 5.1 Codex model, with no other models or services impacted.
Impact
4.5% error rate for users of the GPT 5.1 Codex model in Copilot, leading to degraded performance. No revenue impact or broader service outage was reported.
Trigger
An issue with the upstream model provider for the GPT 5.1 Codex model, which caused the degradation in service.
Detection
GitHub became aware of the incident through internal monitoring and user reports of degraded performance for Copilot at 10:02 UTC.
Resolution
The external model provider implemented a mitigation to resolve the issue. GitHub worked with the provider to improve resiliency and continued monitoring for stability. The service was fully restored by 11:21 UTC.
Root Cause
The root cause was an issue with the upstream model provider's GPT 5.1 Codex model, which led to degraded performance and a 4.5% error rate for users.