Personally been using it with Retrieval tool after uploading files to the assistant and prompting for analysis/responses
For me it’s when I’m submitting a function most of the time, and sometimes it can even happen on simple messages, yesterday it was already having a more difficult time with the queue but today far more.
But i tried the assistent later tonight again and now it’s much better then a couple of hours ago for some reason.
Thanks for chiming in – I was attempting to use the Assistants + Retrieval Tool with scientific articles in PDF format for analysis. It kept hanging on the same 2 PDF files (out of ~8 I was analyzing) but it seems to be working again as of a few minutes ago.
The Assistant was deleted / re-created for each file and I was uploading / removing the file after each analysis – so there was only ever a one-to-one ratio for the Assistant and the file it was analyzing.
Also, I was unable to use the Python openAI package to delete the file if the run changed to “expired”, which was strange. I didn’t investigate further – but I noticed for files with “expired” runs, the file persisted when I checked via the web interface. All other files that didn’t have expired runs were successfully deleted by the script, as expected.
Dont have the issue with gpt-3 but its true that the gpt 4 is kinda slow when paired with the function usage today.
For us it’s function calls. Looks it’s getting better in the day. Has been good for the last couple of hours.
This should be resolved now! Please let me know if you are still seeing this issue.
I confirm that now it works fine and all the runs end with completed status ![]()
Unpinning this as it seems to be resolved.
Logan will most likely close this topic in a few hours to a day so that if the issue reoccurs it can be tracked by starting a new topic.
2 posts were split to a new topic: Assistant run remains queued for a long time - II