Resolved
This incident has been resolved.
Monitoring
Service error rates and latencies for GPT-4o-mini fine-tunes have returned to normal. We will monitoring these services closely to ensure they continue to operate as expected.
Identified
Error rate is returning to 0% but we are still experiencing high variance in request latency. Continuing to monitor.
Identified
A mitigation has been deployed, error rates are stabilizing but still elevated.
Identified
We are continuing to work on a fix. The inference services for 4o-mini fine-tuned models (only 4o-mini) are still experiencing a partial outage.
Identified
We are currently experiencing very high error rates for fine-tuned gpt-4o-mini models.