Skip to main content
POST
/
fine-tune
/
jobs
curl -X POST https://api.u1.archetypeai.app/v0.5/fine-tune/jobs \
  -H "Authorization: Bearer $ATAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "my-finetune-run",
    "config": {
      "model": "Newton::c2_4_7b_251215a172f6d7",
      "datasets": [
        {
          "name": "training-set",
          "split": "train",
          "file_ids": ["file-abc123"]
        },
        {
          "name": "eval-set",
          "split": "eval",
          "file_ids": ["file-def456"]
        }
      ]
    }
  }'

{
  "name": "my-finetune-run",
  "id": "ft-job-my-finetune-run-3de60a5c",
  "config": {
    "model": "Newton::c2_4_7b_251215a172f6d7",
    "datasets": [
      {
        "name": "training-set",
        "split": "train",
        "file_ids": ["training-set.jsonl"]
      },
      {
        "name": "eval-set",
        "split": "eval",
        "file_ids": ["eval-set.jsonl"]
      }
    ]
  },
  "uuid": "job_0n5e9s0w2s9kkrvr8p9fq7sk8w",
  "org_uuid": "org_2wh4zfagp480wbt3nhm8tcqw6z",
  "status": "REGISTERED",
  "created_at": "2026-04-27T14:32:18.421000Z",
  "updated_at": "2026-04-27T14:32:18.421000Z",
  "inference_config_path": null
}

​
Overview

Create a new fine-tune job for the authenticated organization. The job runs the specified model against one or more datasets, partitioned into train and eval splits. Once created, a job is REGISTERED and progresses through STARTING → RUNNING → FINALIZING → COMPLETED (or STOPPED, CANCELLED, or FAILED). Hyperparameters (epochs, learning rate, batch size) are selected automatically from dataset size — the only training-time fields you set are model and datasets. For the dataset file format, see Fine-Tuning and the labeling guide.

​
Request Body

​
name
string
required
Human-readable name for the job
​
id
string
Optional client-supplied identifier for the job (1–64 characters). If omitted, an ID will be generated by the server.
​
config
object
required
Configuration for the fine-tune job

​
Response

​
name
string
Name of the job
​
id
string
Job identifier
​
uuid
string
Server-generated UUID for the job
​
org_uuid
string
UUID of the organization that owns the job
​
status
string
Current status of the job. One of UNKNOWN, REGISTERED, STARTING, RUNNING, FINALIZING, COMPLETED, STOPPING, STOPPED, CANCELLED, FAILED.
​
config
object
Resolved configuration for the job, including the selected model and datasets
​
created_at
string
ISO 8601 timestamp when the job was created
​
updated_at
string
ISO 8601 timestamp when the job was last updated
​
inference_config_path
string
Path to the inference configuration produced by the job. Populated once the job has produced inference artifacts.
curl -X POST https://api.u1.archetypeai.app/v0.5/fine-tune/jobs \
  -H "Authorization: Bearer $ATAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "my-finetune-run",
    "config": {
      "model": "Newton::c2_4_7b_251215a172f6d7",
      "datasets": [
        {
          "name": "training-set",
          "split": "train",
          "file_ids": ["file-abc123"]
        },
        {
          "name": "eval-set",
          "split": "eval",
          "file_ids": ["file-def456"]
        }
      ]
    }
  }'

{
  "name": "my-finetune-run",
  "id": "ft-job-my-finetune-run-3de60a5c",
  "config": {
    "model": "Newton::c2_4_7b_251215a172f6d7",
    "datasets": [
      {
        "name": "training-set",
        "split": "train",
        "file_ids": ["training-set.jsonl"]
      },
      {
        "name": "eval-set",
        "split": "eval",
        "file_ids": ["eval-set.jsonl"]
      }
    ]
  },
  "uuid": "job_0n5e9s0w2s9kkrvr8p9fq7sk8w",
  "org_uuid": "org_2wh4zfagp480wbt3nhm8tcqw6z",
  "status": "REGISTERED",
  "created_at": "2026-04-27T14:32:18.421000Z",
  "updated_at": "2026-04-27T14:32:18.421000Z",
  "inference_config_path": null
}