OpenAI

OpenAI Fine Tuning API

The Fine Tuning API from OpenAI — 11 operation(s) for fine tuning.

Documentation GitHub OpenAPI

Business capability

Artificial Intelligence Management BC-610.60

Operations 13

Each operation below carries the questions people ask an LLM about it and the instructions they give an agent to run it. Generated by API Evangelist overlay

POST /fine_tuning/alpha/graders/run Run a grader against a model sample · Run a grader #

Ask an LLM

“How do I score a single model output with a grader before fine-tuning?”

“Can I test a grader against a sample and a reference item to see the score it gives?”

Tell an agent

Run grader {grader} on the model sample {model_sample}.

Score {model_sample} with grader {grader} using reference item {item}.

POST /fine_tuning/alpha/graders/validate Check that a grader definition is valid · Validate a grader #

Ask an LLM

“Is my grader configuration valid before I use it in a fine-tuning job?”

“Can I check a grader definition for errors without running it?”

Tell an agent

Validate grader definition {grader}.

Check whether {grader} is a well-formed grader without scoring anything.

GET /fine_tuning/checkpoints/{fine_tuned_model_checkpoint}/permissions List who can use a fine-tuned checkpoint · **NOTE:** This endpoint requires an admin API key. #

Ask an LLM

“Which projects have access to one of my fine-tuned model checkpoints?”

“Can I check whether a specific project was granted a checkpoint?”

Tell an agent

List the sharing permissions on checkpoint {fine_tuned_model_checkpoint}.

Show whether project {project_id} can use checkpoint {fine_tuned_model_checkpoint}.

POST /fine_tuning/checkpoints/{fine_tuned_model_checkpoint}/permissions Share a fine-tuned checkpoint with other projects · **NOTE:** Calling this endpoint requires an admin API key. #

Ask an LLM

“How do I share a fine-tuned model with other projects in my organization?”

“Do I need an admin key to grant a project access to a checkpoint?”

Tell an agent

Share checkpoint {fine_tuned_model_checkpoint} with projects {project_ids}.

Grant projects {project_ids} permission to use fine-tuned checkpoint {fine_tuned_model_checkpoint}.

DELETE /fine_tuning/checkpoints/{fine_tuned_model_checkpoint}/permissions/{permission_id} Revoke a project's access to a checkpoint · **NOTE:** This endpoint requires an admin API key. #

Ask an LLM

“How do I stop another project from using my fine-tuned checkpoint?”

“Can I remove a single sharing permission from a checkpoint?”

Tell an agent destructive · confirm first

Delete permission {permission_id} from checkpoint {fine_tuned_model_checkpoint}.

Revoke sharing grant {permission_id} on fine-tuned checkpoint {fine_tuned_model_checkpoint}.

POST /fine_tuning/jobs Start a fine-tuning job to train a custom model · Creates a fine-tuning job which begins the process of creating a new model from… #

Ask an LLM

“How do I fine-tune a model on my own training file?”

“Can I supply a validation file and custom hyperparameters when I start fine-tuning?”

Tell an agent

Fine-tune {model} on training file {training_file}.

Start a fine-tuning job on {model} with {training_file} and validate against {validation_file}.

GET /fine_tuning/jobs List my organization's fine-tuning jobs · List your organization's fine-tuning jobs #

Ask an LLM

“Which fine-tuning jobs has my organization run?”

“Can I filter fine-tuning jobs by the metadata I tagged them with?”

Tell an agent

List all our fine-tuning jobs.

Show fine-tuning jobs tagged with metadata {metadata}.

GET /fine_tuning/jobs/{fine_tuning_job_id} Get the status of a fine-tuning job · Get info about a fine-tuning job. Learn more about fine-tuning #

Ask an LLM

“Is my fine-tuning job finished, and what is the new model called?”

“Can I look up one fine-tuning job by its ID?”

Tell an agent

Show fine-tuning job {fine_tuning_job_id}.

Check whether fine-tuning job {fine_tuning_job_id} has succeeded.

POST /fine_tuning/jobs/{fine_tuning_job_id}/cancel Cancel a fine-tuning job · Immediately cancel a fine-tune job #

Ask an LLM

“How do I stop a fine-tuning job for good?”

“Can I immediately cancel a fine-tune I started by mistake?”

Tell an agent destructive · confirm first

Cancel fine-tuning job {fine_tuning_job_id}.

Abort fine-tune {fine_tuning_job_id} right now.

GET /fine_tuning/jobs/{fine_tuning_job_id}/checkpoints List the checkpoints a fine-tuning job produced · List checkpoints for a fine-tuning job #

Ask an LLM

“Which intermediate checkpoints did my fine-tuning job save?”

“Can I pick an earlier checkpoint from a fine-tuning run instead of the final model?”

Tell an agent

List the checkpoints for fine-tuning job {fine_tuning_job_id}.

Show {limit} saved checkpoints from job {fine_tuning_job_id}.

GET /fine_tuning/jobs/{fine_tuning_job_id}/events Show progress events for a fine-tuning job · Get status updates for a fine-tuning job #

Ask an LLM

“What status updates has my fine-tuning job logged so far?”

“Where can I follow the training progress messages of a fine-tune?”

Tell an agent

Show the events for fine-tuning job {fine_tuning_job_id}.

Get the latest {limit} progress updates from fine-tune {fine_tuning_job_id}.

POST /fine_tuning/jobs/{fine_tuning_job_id}/pause Pause a running fine-tuning job · Pause a fine-tune job #

Ask an LLM

“Can I temporarily pause a fine-tuning job without cancelling it?”

“Is there a way to hold a fine-tune and pick it up later?”

Tell an agent

Pause fine-tuning job {fine_tuning_job_id}.

Put fine-tune {fine_tuning_job_id} on hold for now.

POST /fine_tuning/jobs/{fine_tuning_job_id}/resume Resume a paused fine-tuning job · Resume a fine-tune job #

Ask an LLM

“How do I restart a fine-tuning job I paused earlier?”

“Can a paused fine-tune continue training from where it stopped?”

Tell an agent

Resume fine-tuning job {fine_tuning_job_id}.

Continue the paused fine-tune {fine_tuning_job_id}.

OpenAI Fine Tuning API

Operations 13

Documentation

Specifications

Schemas & Data

Other Resources

Work with this as data

MCP server

Tools for apis

Call it yourself

Get an API key

OpenAPI Specification