Fine-tuning

The Fine-tuning API proxies fine-tuning job management to the configured provider. All endpoints follow the OpenAI fine-tuning API contract. The correct operation is determined automatically by the HTTP method and the presence of a :jobId path parameter.

Fine-tuning requires uploading a training file first. Use POST /v1/files with purpose=fine-tune to prepare your data.

Endpoint routing

All fine-tuning requests are handled at /v1/fine_tuning/jobs/:jobId?/:cancel?. The gateway dispatches to the correct operation based on the method and path:

Method	Path	Operation
`GET`	`/v1/fine_tuning/jobs`	List fine-tuning jobs
`POST`	`/v1/fine_tuning/jobs`	Create a fine-tuning job
`GET`	`/v1/fine_tuning/jobs/:id`	Retrieve a fine-tuning job
`POST`	`/v1/fine_tuning/jobs/:id/cancel`	Cancel a fine-tuning job

List fine-tuning jobs

GET /v1/fine_tuning/jobs Returns a list of fine-tuning jobs for the configured provider.

Request

string

required

The provider to route the request to.

string

required

Your Portkey API key or provider API key.

number

default:"20"

Maximum number of jobs to return.

string

Cursor for pagination. Pass the id of the last job from the previous page.

Response

string

Always list.

array

Array of fine-tuning job objects.

boolean

Whether additional jobs are available.

curl https://your-gateway.example.com/v1/fine_tuning/jobs \
  -H "x-portkey-api-key: YOUR_API_KEY" \
  -H "x-portkey-provider: openai"

from openai import OpenAI

client = OpenAI(
    base_url="https://your-gateway.example.com/v1",
    api_key="YOUR_API_KEY",
    default_headers={
        "x-portkey-provider": "openai",
    },
)

jobs = client.fine_tuning.jobs.list()
for job in jobs.data:
    print(job.id, job.status)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://your-gateway.example.com/v1",
  apiKey: "YOUR_API_KEY",
  defaultHeaders: {
    "x-portkey-provider": "openai",
  },
});

const jobs = await client.fineTuning.jobs.list();
for (const job of jobs.data) {
  console.log(job.id, job.status);
}

Create a fine-tuning job

POST /v1/fine_tuning/jobs Creates a new fine-tuning job. The request body is forwarded directly to the provider.

Request

string

required

The provider to route the request to.

string

required

Your Portkey API key or provider API key.

string

required

The ID of the uploaded training file (must have purpose=fine-tune).

string

required

The base model to fine-tune. For example, gpt-4o-mini-2024-07-18.

string

The ID of an optional validation file.

object

Show properties

number | string

Number of training epochs. Use auto to let the provider decide.

number | string

Training batch size. Use auto for provider default.

number | string

Multiplier applied to the base learning rate. Use auto for provider default.

string

A string of up to 18 characters appended to the fine-tuned model name.

object

Optional key-value metadata to attach to the job.

Response

string

Unique identifier for the fine-tuning job.

string

Always fine_tuning.job.

string

The base model being fine-tuned.

string

The name of the fine-tuned model once training completes. null while in progress.

string

Current job status. Values: validating_files, queued, running, succeeded, failed, cancelled.

number

Unix timestamp when the job was created.

number

Number of tokens used in training, or null if not yet available.

curl https://your-gateway.example.com/v1/fine_tuning/jobs \
  -X POST \
  -H "Content-Type: application/json" \
  -H "x-portkey-api-key: YOUR_API_KEY" \
  -H "x-portkey-provider: openai" \
  -d '{
    "training_file": "file-abc123",
    "model": "gpt-4o-mini-2024-07-18"
  }'

job = client.fine_tuning.jobs.create(
    training_file="file-abc123",
    model="gpt-4o-mini-2024-07-18",
)
print(job.id, job.status)

const job = await client.fineTuning.jobs.create({
  training_file: "file-abc123",
  model: "gpt-4o-mini-2024-07-18",
});
console.log(job.id, job.status);

Retrieve a fine-tuning job

GET /v1/fine_tuning/jobs/:id Retrieves details for a specific fine-tuning job. Use this to poll the status of an in-progress job.

Request

string

required

The ID of the fine-tuning job to retrieve.

string

required

The provider to route the request to.

string

required

Your Portkey API key or provider API key.

Response

Returns a single fine-tuning job object. See Create a fine-tuning job for the full field list.

curl https://your-gateway.example.com/v1/fine_tuning/jobs/ftjob-abc123 \
  -H "x-portkey-api-key: YOUR_API_KEY" \
  -H "x-portkey-provider: openai"

job = client.fine_tuning.jobs.retrieve("ftjob-abc123")
print(job.status, job.fine_tuned_model)

const job = await client.fineTuning.jobs.retrieve("ftjob-abc123");
console.log(job.status, job.fine_tuned_model);

Cancel a fine-tuning job

POST /v1/fine_tuning/jobs/:id/cancel Cancels a fine-tuning job that is queued or running.

Once a job is cancelled it cannot be resumed. Any compute costs incurred up to the cancellation point may still apply.

Request

string

required

The ID of the fine-tuning job to cancel.

string

required

The provider to route the request to.

string

required

Your Portkey API key or provider API key.

Response

Returns the updated fine-tuning job object with status: cancelled.

curl https://your-gateway.example.com/v1/fine_tuning/jobs/ftjob-abc123/cancel \
  -X POST \
  -H "x-portkey-api-key: YOUR_API_KEY" \
  -H "x-portkey-provider: openai"

job = client.fine_tuning.jobs.cancel("ftjob-abc123")
print(job.status)  # cancelled

const job = await client.fineTuning.jobs.cancel("ftjob-abc123");
console.log(job.status); // cancelled

Overview

Chat

Multimodal

Files & Batches

Other

Endpoint routing

List fine-tuning jobs

Request

Response

Create a fine-tuning job

Request

Response

Retrieve a fine-tuning job

Request

Response

Cancel a fine-tuning job

Request

Response

​Endpoint routing

​List fine-tuning jobs

​Request

​Response

​Create a fine-tuning job

​Request

​Response

​Retrieve a fine-tuning job

​Request

​Response

​Cancel a fine-tuning job

​Request

​Response

Endpoint routing

List fine-tuning jobs

Request

Response

Create a fine-tuning job

Request

Response

Retrieve a fine-tuning job

Request

Response

Cancel a fine-tuning job

Request

Response