Breaking rest
From January 6, 2027, no OpenAI organization can create new fine-tuning jobs; existing fine-tuned models keep serving inference until their base model shuts down
What changes
Notice date May 7, 2026. OpenAI's deprecations page lists self-serve fine-tuning job creation with a January 6, 2027 date and status Restricted access, pointing to its fine-tuning timeline. That timeline stopped job creation for organizations that had never run fine-tuning (May 7, 2026) and for those with no fine-tuned model inference in the past 60 days (July 2, 2026), and says that on January 6, 2027 active existing customers will no longer be able to create new fine-tuning jobs. Inference on fine-tuned models will be disabled only when the underlying base model is deprecated. The Supervised fine-tuning guide says eligible existing customers can create training jobs until January 6, 2027, unless the base model shuts down earlier. OpenAI names no replacement for training new fine-tuned models.
Breaking: Calls fail after this date.
What to do
Run any training you still need before January 6, 2027, and remove automated job creation that would start failing then. Existing fine-tuned models keep serving inference until their base model's shutdown date on OpenAI's deprecations page, so plan to move those workloads to a current base model by that date; OpenAI names no replacement for creating new fine-tunes.
Strings to search for
fine_tuning.jobs.createfineTuning.jobs.create/v1/fine_tuning/jobs