Timeout after 300 seconds with any error message. Could it be rate limiting?

[jackien1]

Just a timeout without any error message. Happens pretty often with the new gpt-4.1 model. How can I reduce this?

I'm Tier 5 on OpenAI so I don't think its rate limiting?

[jackien1]

I checked forums and saw some mentions of errors in regards to the assistants api. I'm using the responses api so maybe I'll try switching to chat completions for now.

[rm-openai]

Are you using long context/is your input very long? I can investigate. Can you also post the full stack trace/error messages that you see?

[jackien1]

My context is around 2-4k tokens. It turned out to be a nested function call input that I flattened which worked. gpt-4o worked flawlessly but gpt-4.1 had the 300 second timeout. Also, got another related timeout issue. Function call executions (not input) that take longer than 300 seconds cause the agent to stop running. No errors but no follow up output.

[rm-openai]

Can you share some sample code to reproduce the issue? happy to take a look

[jackien1]

Can you share some sample code to reproduce the issue? happy to take a look

Is there private way of sending it via GitHub? I can't send any sample code since the erroring function is super long and proprietary.

[rm-openai]

You can email it to me - though please remove any sensitive or proprietary stuff you don't want to share. rm at OpenAI dot com

[jackien1]

You can email it to me - though please remove any sensitive or proprietary stuff you don't want to share. rm at OpenAI dot com

Sent an email sorry for the delay!

[fitzjalen]

Same problem - long context + reasoning model = timeout error