OpenAI API timeout (no response in time)
Is there an outage right now?
OpenAI API is currently reporting issues. This error may be caused by a wider outage.
Updated: Jul 21, 2026, 8:00 PM UTC
View OpenAI API status →What this error means
The response does not return within the time limit: long generations, network, overload, or too-short client timeout.
Common causes
- •Generation too long to finish in time
- •Unstable network
- •OpenAI-side overload
- •Client timeout too short
- •Proxy/gateway timeout
How to fix
- 1.Use streaming (stream=true)
- 2.Reduce max_tokens / input
- 3.Increase the client timeout
- 4.Retry with exponential backoff
- 5.Check network/proxy settings
When waiting helps (and when it won't)
Improves over time if overload-related. If persistent, switch to streaming and review timeouts.
Alternatives when OpenAI API is down
Related errors
FAQ
What is Timeout / gateway timeout?
The response does not return within the time limit: long generations, network, overload, or too-short client timeout.
How do I fix Timeout / gateway timeout?
Use streaming (stream=true) / Reduce max_tokens / input / Increase the client timeout / Retry with exponential backoff / Check network/proxy settings