Closed bhaskoro-muthohar closed 1 year ago
Hi, I’ve seen this happen with GPT4. I’m curious what model you were using?
I think if the prompt is over some complexity threshold then the model either gets “confused” or just takes too long causing the OpenAI API gateway to abort the request. I’m guessing about this.
I’ve found that when I make the prompt more concise then the timeout is usually fixed. I have no idea what’s happening of course, but I recommend trying to shorten or somehow make your prompt less confusing when this happens.
Is there any way to fix the gateway timeout error?