gpt-image-2.5-sunburst / gpt-image-2.5-flare between 15:30 and 16:00 (UTC+8) on 9/17 have been resolved
During that window some requests returned Upstream model service error. Try again later. The cause was backend resource fluctuation. It was not a request parameter error, and not a prompt or content-safety block. The issue is fixed, so no code, parameter or prompt changes are needed; simply resubmit any tasks that failed at the time.
If you see this error again, retry after a few seconds first. If it keeps failing, contact support with the request time and the exact error text.
We apologize for the disruption; we are monitoring continuously.
← Back to Live Updates · 📚 Monthly archive