The run status remains “queued”.
It started to happen yesterday. Does anyone see the same?
Any idea?
8 Likes
I’m facing this now… It was working few minutes ago but now the Run remains in the queued status for ever.
3 Likes
I have the same issue. Runs are getting stuck in queued and then they just expire. It seems that it usually happens after submitting tool outputs (at least for me). Also, it is impossible to cancel those runs. They are just stuck in cancelling until expired.
1 Like
I am encountering a similar issue that commenced approximately yesterday. I am inquiring whether it may be related to the Assistant’s API or if it pertains to the submission of tool outputs. Personally, I observe this issue occurring consistently, particularly when using the submitting tool outputs; however, it also manifests with simple messages.
3 Likes
I am encountering the same problem. I noticed this issue only today though.
1 Like
Yes, same here. My code is ten days old and was working fine. Now out of the sudden I get intermittent ‘queued’ states that don’t seem to go away or at least I never waited long enough. Anyhow, whenever this happens, the bot “hangs” which isn’t good.
I face the same problem since yesterday.
I pinned this because it has a few post noting the same problem in a very short time frame, less than 30 minutes, and similar post are starting to appear.
Usually a pinned topic when it is a real problem will get more replies confirming the issue and OpenAI then tends to get if fixed faster.
If you are seeing this problem just note it here instead of starting a new topic.
Had this this morning but only with GPT-4-1106 preview
Same, I see GET status: queued, after that, I see status: expired. It was working fine yesterday.
Likewise experiencing this issue on multiple assistant deployments. It has happened before but now it is happening about 30% of the time, including with very small context windows (even in first ‘hello’ messages for assistants with short instructions)
Also experiencing this issue on gpt-4-1106. Have also had a solution built for a few weeks now that’s been working 100%.
When it does run, it seems to have issues processing the files I upload.
The reality of using and relying on a beta 
1 Like
Not only gpt-4-1106-preview. I have the same problem with gpt-3.5-turbo-1106
1 Like
Experiencing issue as well, runs staying queued for an inordinate amount of time before expiring. I thought it might be related to the file upload issues but its happening with text-only prompts as well. Using assistants API, both 4 and 3.5 1106 models.
1 Like
same issue. GPT-4 34k and 128k are doesnt work
Just started my first tests using the Assistants API this morning, and i am seeing the same thing. For me it seems to be happening when I submit a message that would require a function call. Normal chit-chat type stuff gets responses. Overall chat completions seem to be faster and far more reliable than Assistants. Bummer.
Assistants API has been great outside of today. I’d recommend waiting for resolution with this issue and trying again.
2 Likes
Yeah, i’ll go back to this tomorrow or soon. Of course the day I try to kick the tires the whole thing is off the rails. Thanks!
Hey folks, thank you for flagging, can anyone chime in on if there are specific tools that are causing this? I have all the tools enabled and I am getting a response in my tests.
2 Likes
For me it enters endless queue status when there would be a need for a function call. I never get the status of ‘requires_action’. I do get simple chat responses though, it is a bit slower than using chat completions from my early stages of testing. There are some benefits of using the assistant api for sure but if it is slower and more bug-ridden than the chat completions, it really isn’t worth using at all. 