gpt-image-2-vip saw about 15 minutes of concurrency saturation from 13:46 to 14:00 (UTC+8) and has been back to normal since 14:00 sharp
Cause: the saturation came from insufficient concurrency in the account pool behind gpt-image-2-vip, not from the gateway or request parameters. This model runs on a reverse channel, which is by nature an ongoing contest with the provider’s platform risk controls: pool capacity fluctuates with risk-control actions, so brief saturation cannot be ruled out entirely. No code or parameter changes are needed after recovery.
We suggest wiring the official gpt-image-2 into your pipeline as a fallback: both use the same images API with compatible parameters, so requests can fail over automatically the next time gpt-image-2-vip fluctuates. See the gpt-image-2 integration guide.
Thank you for your patience; we are monitoring continuously.
← Back to Live Updates · 📚 Monthly archive