I’m defining my own function_tool and output_type. the span link: https://platform.openai.com/logs/resp_67db62f56abc81928070d62b579265cb0972ffc967b4f996
The entire run takes about 100 second. I figure it’s because of my specified output type. But why is it taking so long with a output_type. Thanks
Same here, more than 40 seconds for first token with short prompt and short instructions.
Edit: I checked and yesterday the same request was taking around 2 seconds, now it’s 40-50 seconds, so I think it’s an OpenAI issue
I came here to the forum’s to see if anyone else was noticing poor performance. I keep a log on timeouts and failed responses and it’s getting worse. February was poor but March is getting worse not better.
As someone who is re-selling this is a poor level of customer service and I find myself at a loss for what to tell my user base.
I’d pin point the degradation in service to around the end of January. Service was fine up until then.