Fine-tuning job stuck in queued state

API Bugs
1

Hi, it seems to be happening again for me. I’ve been trying to fine-tune a model since Friday, and I’ve triggered different runs, but all of them stay queued for several hours before I cancel them. The current job has ID ftjob-pay4lQOKW7rtTx9vGvOTy82T.

Information

  • I already changed the model from gpt-4.1-mini-2025-04-14 to gpt-4.1-nano-2025-04-14, but I got the same issue.

  • I already canceled jobs that were queued and retried them.

  • It is the only job running in my org.

  • The training file has fewer than 500 examples, and they are all short phrases, so it should not be related to token size.

2

Yesterday, it started running again. It is still getting stuck in the queue for 6–7 hours before starting the fine-tuning job, even though I am running only one job at a time in my org, but I suppose the bug can be closed.