OpenAI HTTP request is failing a lot
# supporta
Adrian
10/07/2025, 5:24 PMLatency, using a reasoning model in OpenAi can take 10-20 seconds for a response so if your agent isn’t expecting that it might call it again too soon, using LLM is not instant and adds a LOT of latency, especially for specific non real time prompts
k
Kizzz
10/07/2025, 5:37 PMSo nothing can be done to increase success?
a
Adrian
10/07/2025, 5:37 PMWhat are you using OpenAi for?
k
Kizzz
10/07/2025, 7:10 PMThats my model for my assistant I am not sure what you mean. It uses my prompt to generate responses.
a
Adrian
10/07/2025, 7:11 PMi see, i thought you were doing an external api
Adrian
10/07/2025, 7:12 PMi'm not sure why it's calling twice then
k
Kizzz
10/07/2025, 7:49 PMSometimes it calls twice and others it works the first time im not sure what I can do
Kizzz
10/07/2025, 8:07 PMI think maybe its retrying the request too early.
Kizzz
10/07/2025, 8:09 PMBecause I look at normal time between request and output and for my assistant its usually 1.2-1.4 seconds and i noticed that it retrys the request at 1.5seconds. Im thinking maybe if it didnt retry the first request is probably almost done.
Kizzz
10/07/2025, 8:09 PMIs there a way to extend the time it waits before sending another model request? Instead of 1.5 seconds which is how long it seems to wait maybe set that too like 2 seconds or longer
Kizzz
10/08/2025, 2:29 PMDoes anyone know if that is possible?
4 Views