Implementation:Openai Openai node FineTuning Jobs Resource

Knowledge Sources	Openai_Openai_node
Domains	SDK, Fine-Tuning, Model Training
Last Updated	2026-02-15 12:00 GMT

Overview

The Jobs resource class provides methods for creating, managing, and monitoring fine-tuning jobs through the OpenAI API.

Description

The Jobs class extends APIResource and serves as the primary interface for the OpenAI fine-tuning jobs endpoint (/fine_tuning/jobs). It enables developers to create fine-tuning jobs that train custom models from uploaded datasets, retrieve job status, list all jobs in an organization, cancel running jobs, and monitor job events.

The resource supports three fine-tuning methods: supervised, DPO (Direct Preference Optimization), and reinforcement. Each method has its own configuration options accessible through the method parameter. The class also provides lifecycle management with pause and resume methods for controlling running jobs. Pagination is handled through cursor-based pagination via CursorPage.

A nested checkpoints sub-resource is available for accessing fine-tuning job checkpoints. The class defines comprehensive TypeScript types for job objects, events, parameters, and Weights & Biases integration.

Usage

Use this resource when you need to fine-tune OpenAI models on custom training data. This includes creating new fine-tuning jobs with specific model and dataset configurations, monitoring job progress through events, managing job lifecycle (pause, resume, cancel), and listing historical jobs for your organization.

Code Reference

Source Location

• Repository: openai-node
• File: src/resources/fine-tuning/jobs/jobs.ts

Signature

export class Jobs extends APIResource {
  checkpoints: CheckpointsAPI.Checkpoints;

  create(body: JobCreateParams, options?: RequestOptions): APIPromise<FineTuningJob>;
  retrieve(fineTuningJobID: string, options?: RequestOptions): APIPromise<FineTuningJob>;
  list(query?: JobListParams | null, options?: RequestOptions): PagePromise<FineTuningJobsPage, FineTuningJob>;
  cancel(fineTuningJobID: string, options?: RequestOptions): APIPromise<FineTuningJob>;
  listEvents(fineTuningJobID: string, query?: JobListEventsParams | null, options?: RequestOptions): PagePromise<FineTuningJobEventsPage, FineTuningJobEvent>;
  pause(fineTuningJobID: string, options?: RequestOptions): APIPromise<FineTuningJob>;
  resume(fineTuningJobID: string, options?: RequestOptions): APIPromise<FineTuningJob>;
}

Import

import OpenAI from 'openai';

I/O Contract

Inputs

Name	Type	Required	Description
model	'babbage-002' | 'davinci-002' | 'gpt-3.5-turbo' | 'gpt-4o-mini'	Yes	The name of the model to fine-tune
training_file	string	Yes	The ID of an uploaded JSONL file containing training data
validation_file	null	No	The ID of an uploaded file containing validation data
method	JobCreateParams.Method	No	Fine-tuning method: supervised, dpo, or reinforcement
hyperparameters	JobCreateParams.Hyperparameters	No	Deprecated in favor of method; controls batch_size, learning_rate_multiplier, n_epochs
suffix	null	No	A string up to 64 characters added to the fine-tuned model name
seed	null	No	Controls reproducibility of the job
metadata	null	No	Set of 16 key-value pairs for storing additional information
integrations	null	No	Integrations to enable (currently only Weights and Biases)

Outputs

Name	Type	Description
id	string	The object identifier for the fine-tuning job
status	'queued' | 'running' | 'succeeded' | 'failed' | 'cancelled'	Current status of the job
model	string	The base model being fine-tuned
fine_tuned_model	null	The name of the resulting fine-tuned model (null while running)
training_file	string	File ID used for training
created_at	number	Unix timestamp of job creation
finished_at	null	Unix timestamp of job completion
trained_tokens	null	Total billable tokens processed
error	null	Error details if the job failed
result_files	Array<string>	Compiled results file IDs
hyperparameters	FineTuningJob.Hyperparameters	Hyperparameters used (for supervised jobs)
method	FineTuningJob.Method	The fine-tuning method used

Usage Examples

const client = new OpenAI();

// Create a fine-tuning job
const fineTuningJob = await client.fineTuning.jobs.create({
  model: 'gpt-4o-mini',
  training_file: 'file-abc123',
});

// Retrieve a fine-tuning job
const job = await client.fineTuning.jobs.retrieve('ft-AF1WoRqd3aJAHsqc9NY7iL8F');

// List all fine-tuning jobs with auto-pagination
for await (const fineTuningJob of client.fineTuning.jobs.list()) {
  console.log(fineTuningJob.id, fineTuningJob.status);
}

// Cancel a running job
const cancelledJob = await client.fineTuning.jobs.cancel('ft-AF1WoRqd3aJAHsqc9NY7iL8F');

// List events for a job
for await (const event of client.fineTuning.jobs.listEvents('ft-AF1WoRqd3aJAHsqc9NY7iL8F')) {
  console.log(event.message);
}

// Pause and resume a job
await client.fineTuning.jobs.pause('ft-AF1WoRqd3aJAHsqc9NY7iL8F');
await client.fineTuning.jobs.resume('ft-AF1WoRqd3aJAHsqc9NY7iL8F');

Key Types

FineTuningJobEvent

interface FineTuningJobEvent {
  id: string;
  created_at: number;
  level: 'info' | 'warn' | 'error';
  message: string;
  object: 'fine_tuning.job.event';
  data?: unknown;
  type?: 'message' | 'metrics';
}

FineTuningJobWandbIntegration

interface FineTuningJobWandbIntegration {
  project: string;
  entity?: string | null;
  name?: string | null;
  tags?: Array<string>;
}

Related Pages

• Environment:Openai_Openai_node_Node_20_Runtime