Am I the only one who can’t fine tune, or is it down again? I’ve been waiting now for 2 days, and job didn’t start … ![]()
Hi!
I have seen a few reports about intermittent issues with the fine-tuning API in the last week and will ping the team.
In the meantime the only workaround is to restart the job until it works.
Hope this gets you unblocked until this issue is resolved for good.
Thx, it’s the second time in a couple of months they’re struggling apparently. It’s painful since I’m tuning on 255,000 rows, 200MB, so it spends a LOT of time simply validating my material, but thx …
Did you get a reply from the team? It’s still hanging on “Waiting …” ![]()
Still waiting for feedback.
As far as I can tell, there are two options: either wait for the job to eventually start, or cancel the current job and start a new one.
Thx, but I’ve cancelled it several times already, and it simply doesn’t start. Same problem they had a month ago … ![]()
I’ll try to reproduce this issue to help raise visibility. Which model are you using, and what modalities are involved?
If you can share a request ID, I can also forward it if needed.
4.1-mini, my job id is ftjob-KBl8nXbGqEh5KFZsAkvtawV5
Hello, We’ve confirmed that this is affecting some fine-tuning jobs waiting in queue, including jobs on gpt-4.1-mini. The jobs are being accepted and validated, but some are taking longer than usual to start training due to elevated queue volume. We’ve increased processing throughput and expect the queue to clear gradually over the next several hours, though some jobs may still take up to 24 hours. For large training files, it’s best to keep one validated job queued rather than repeatedly cancelling and recreating it.
I started a job about 10 hours ago, it’s still in the “waiting” state. Should I restart it, or? This is my job ID “ftjob-JVOl0HdHVUukUACZL3QRJsRa”
Come on, seriously guys? I’ve been trying to start my job since Sunday now. I recreated the job about 10 hours ago, and nothing still. What’s happening? This used to take 3/5 minutes, now it’s talking 3/5 DAYS, maybe more? Who knows?
Please …?
Hello, this issue has now been fixed.
At the moment, only jobs submitted within the last 9 hours may still appear as queued or in progress. Jobs submitted earlier than that should now be completed.
We’re marking this as resolved for now. If you’re still experiencing issues, please contact Support through help.openai.com and include your job ID so we can take a closer look.