I have been trying to fine-tune gpt-4.1-2025-04-14 for a few days and no matter I try it doesn’t seem to ever move out of the queued state. I’ve waited 12+ hours and nothing ever happens. I can take the same training and validation data sets and they work fine on gpt-4.1-mini-2025-04-14. I have also tried very simple files (simple QA training set) with only 30 training samples. The files always validate fine. Then the log says “Files validated, moving job to queued state” and then it just sits there forever.
I see in the forum history similar things have happened, but it was a few years ago, and the issue seemed to resolve itself eventually. But it has been a week at this point.
I’m doing everything via the fine-tuning UI, not the API. Anyone else having these issues or can confirm it works fine for them?
I haven’t needed to wait this long I think since 2023 or something. It’s been stuck for hours, in “waiting state”. Can somebody from OpenAI at least maybe drop an announcement or something related to this? When might we expect it to get back up again?
Unfortunately, now similar fine-tuning jobs have been stuck in the “Waiting…” state for several hours, which is anomalous. This has happened with several jobs in a row. I typically see this state for less than a minute, between “Validating files” and “Fine-tuning”.
I hope someone will see this and get it working again, and hopefully improve the overall trustworthiness of this service.
When you say you got 1 through, did a stuck one finally complete or did you start a fresh one and it went through? I’m asking so I can gauge if I should cancel one that’s been stuck for 18+ hours and start new or not.
There was an issue that the engineers identified and fixed. Please restart any impacted jobs, and let us know if you are still seeing issues. We apologize for the inconvenience.
Has fine-tuning worked for anyone after the reported fix? We’re still experiencing the same issue: none of our fine-tuning jobs have moved beyond “Files validated, moving job to queued state” since March 23rd. We tried running some jobs again this morning, but we’re still seeing the same problem
We’re fine-tuning from the gpt-4.1-2025-04-14 base model.
hi! we have fixed the queue issue. if your job is still failing, would you mind sending the job_id so the team could take a look? thank you for your patience!
Thanks for the update! The two fine-tuning jobs we ran this morning both completed successfully, so it looks like the queue issue may be resolved for us.
That said, they took quite a bit longer than usual - around 10–11 hours, whereas similar jobs have typically taken 2–4 hours in the past. Should we expect turnaround times to improve over the next few days, or is this the new normal for now?
Hi，i have created a job for fining tuning by using gpt-4.1-mini-2025-04-14, and it started from
12:25 PM this morning, and almost 4 hours, it is still in “Files validated, moving job to queued state“
This particular issue has been resolved.
If you are running into the same problem in April 2026 please follow the link below.
Otherwise please create a new topic.
