How to reduce latency when calling the OpenAI API?

Hey folks, we just shipped an updated section of our “Production best practices” guide on Latency and how to reduce its impact on your application. S/o to @shyamal for pushing this forward:


Still slow AF. Takes about 13 seconds to return completions - was about 7 seconds a few days ago, which is still really slow. What’t going on?

It’s no secret “what is going on” @michael10

The number of OpenAI users have increased at the largest rate of any Internet technology in the history of the world.

The last I read, OpenAI went from 1 million users to 100 million in about a months time just recently. Not sure what the user count is now, but OpenAI’s infrastructure is currently fragile and OpenAI is working to “beef it up” to meet the demands.

Hope this helps.


1 Like