ChatGPT Is Down (status.openai.com)

ChatGPT experienced a brief service disruption that affected the Platform Billing page and the Responses API. Users saw the error “Failed to load subscription or selected organization” when accessing billing, and roughly 10% of requests to the Responses API returned HTTP 500 server errors for about 12 minutes. The incident has been fully resolved and all impacted services have recovered. For the AI/ML community, even short outages matter: billing/organization UI failures can block account management and integrations, while intermittent 500s in the Responses API can break pipelines, real-time apps, or batch jobs that assume high availability. The 10% error rate over a 12-minute window suggests a transient backend issue rather than a complete platform failure, but it underscores the importance of robust client-side handling—retries with exponential backoff, idempotency keys, graceful degradation, and monitoring/alerts tied to error-rate thresholds. Teams using the API should verify any critical tasks that ran during the window and confirm retry behaviors; platform operators may want to review incident postmortems to tighten SLAs and reduce blast radius for similar transient faults.