Is text-davinci-003 still down?

text-davinci-003 has been acting up since yesterday. Im accessing it through the API, on status support page it says it was resolved yesterday. But nothing has changed. still getting Error code 500

1 Like

Sorry to inform that I have been using these models all day with no issues but maybe two server-side API timeouts.


Same. Get error all the day. may be 1 success response in 20 request
and i tried to change my api key, make the request in playground, still error

1 Like

same for me. A lot of failed requests today.
And the duration of the request is approximately 15 - 20 times longer than usually

1 Like

you mean you have no issue ? what location are the requests made from in your case ? maybe this has to do with the country

I am facing the same issue, even today. I am located in Slovenia (EU). Can’t get any reply from openai, and this is killing my product. Does someone have any info what is going on?

1 Like

Yes, no issue. I’m am located on the seacoast of amazing Thailand where the weather is fantastic, the food is delicious, and everyone loves to smile.


Today, also no issues at all, to be honest.

Hope this issue clears up for everyone soon.

Note: I think it’s also related to Cloudflare configuration issues and not always directly an OpenAI infrastructure issue, but I could be wrong since I don’t have visibility into the OpenAI Cloudflare configuration.


I’m curious if those that had issues are up and running again. I still seem to be having the same issue reported by many of you. Just trying to figure out if its my end or still the api?

1 Like

Yea I have the same issue. Very unstable to be honest… I have a live app with 5k users and so many times the completion just fails with Davinchi 003… 001 and 002 seem to work ok.

1 Like

No, it’s not better at all. And I can’t find any info if/when they will do something about it.

The discord channel has mentioned that latency (delay in responding) has been addressed in the last hour or so

I don’t see lots of chatter about downtime

The latency issue would look like the API was not responding. Some people had indicated this was in the order of 40 seconds+

But it appears to have been recently addressed

Last night it started to work at cca 22.15 CET, probably to to lower traffic? Today, it is the same situation, it can’t connect to remote server. Just frustrating…