The Mistral AI Fine-Tuning API allows developers to create and manage fine-tuning jobs that customize Mistral models on proprietary datasets. Users can upload training data, configure hyperparameters, and monitor job progress through the API. Fine-tuned models can then be deployed and accessed through the Chat Completions endpoint, enabling domain-specific performance improvements for enterprise applications.
openapi: 3.1.0
info:
title: Mistral AI Fine-Tuning API
description: >-
The Mistral AI Fine-Tuning API allows developers to create and manage
fine-tuning jobs that customize Mistral models on proprietary datasets.
Users can upload training data, configure hyperparameters, and monitor
job progress through the API. Fine-tuned models can then be deployed
and accessed through the Chat Completions endpoint, enabling domain-
specific performance improvements for enterprise applications.
version: '1.0.0'
contact:
name: Mistral AI Support
url: https://docs.mistral.ai
termsOfService: https://mistral.ai/terms
externalDocs:
description: Mistral AI Fine-Tuning Documentation
url: https://docs.mistral.ai/api/endpoint/fine-tuning
servers:
- url: https://api.mistral.ai/v1
description: Mistral AI Production Server
tags:
- name: Fine-Tuning Jobs
description: >-
Endpoints for creating, managing, and monitoring fine-tuning jobs
on Mistral AI models.
security:
- bearerAuth: []
paths:
/fine_tuning/jobs:
post:
operationId: createFineTuningJob
summary: Create fine-tuning job
description: >-
Creates a fine-tuning job which begins the process of creating a new
model from a given dataset. The job runs asynchronously and its
progress can be monitored via the retrieve endpoint.
tags:
- Fine-Tuning Jobs
requestBody:
required: true
content:
application/json:
schema:
$ref: '#/components/schemas/CreateFineTuningJobRequest'
responses:
'200':
description: Successfully created fine-tuning job
content:
application/json:
schema:
$ref: '#/components/schemas/FineTuningJob'
'400':
description: Bad request due to invalid parameters
content:
application/json:
schema:
$ref: '#/components/schemas/Error'
'401':
description: Unauthorized due to missing or invalid API key
content:
application/json:
schema:
$ref: '#/components/schemas/Error'
get:
operationId: listFineTuningJobs
summary: List fine-tuning jobs
description: >-
List fine-tuning jobs for your organization. Returns a paginated
list of jobs with their current status and configuration.
tags:
- Fine-Tuning Jobs
parameters:
- $ref: '#/components/parameters/PageParam'
- $ref: '#/components/parameters/PageSizeParam'
responses:
'200':
description: Successfully retrieved list of fine-tuning jobs
content:
application/json:
schema:
$ref: '#/components/schemas/FineTuningJobList'
'401':
description: Unauthorized due to missing or invalid API key
content:
application/json:
schema:
$ref: '#/components/schemas/Error'
/fine_tuning/jobs/{job_id}:
get:
operationId: getFineTuningJob
summary: Get fine-tuning job
description: >-
Get info about a fine-tuning job including its status, configuration,
training metrics, and resulting model identifier.
tags:
- Fine-Tuning Jobs
parameters:
- $ref: '#/components/parameters/JobIdParam'
responses:
'200':
description: Successfully retrieved fine-tuning job
content:
application/json:
schema:
$ref: '#/components/schemas/FineTuningJob'
'401':
description: Unauthorized due to missing or invalid API key
content:
application/json:
schema:
$ref: '#/components/schemas/Error'
'404':
description: Fine-tuning job not found
content:
application/json:
schema:
$ref: '#/components/schemas/Error'
/fine_tuning/jobs/{job_id}/cancel:
post:
operationId: cancelFineTuningJob
summary: Cancel fine-tuning job
description: >-
Request the cancellation of a fine-tuning job. The job will be
cancelled if it has not already completed or failed.
tags:
- Fine-Tuning Jobs
parameters:
- $ref: '#/components/parameters/JobIdParam'
responses:
'200':
description: Successfully cancelled fine-tuning job
content:
application/json:
schema:
$ref: '#/components/schemas/FineTuningJob'
'401':
description: Unauthorized due to missing or invalid API key
content:
application/json:
schema:
$ref: '#/components/schemas/Error'
'404':
description: Fine-tuning job not found
content:
application/json:
schema:
$ref: '#/components/schemas/Error'
/fine_tuning/jobs/{job_id}/start:
post:
operationId: startFineTuningJob
summary: Start fine-tuning job
description: >-
Start a fine-tuning job that was created with auto_start set to false.
The job must have passed validation before it can be started.
tags:
- Fine-Tuning Jobs
parameters:
- $ref: '#/components/parameters/JobIdParam'
responses:
'200':
description: Successfully started fine-tuning job
content:
application/json:
schema:
$ref: '#/components/schemas/FineTuningJob'
'401':
description: Unauthorized due to missing or invalid API key
content:
application/json:
schema:
$ref: '#/components/schemas/Error'
'404':
description: Fine-tuning job not found
content:
application/json:
schema:
$ref: '#/components/schemas/Error'
components:
securitySchemes:
bearerAuth:
type: http
scheme: bearer
bearerFormat: API Key
parameters:
JobIdParam:
name: job_id
in: path
required: true
description: >-
The ID of the fine-tuning job.
schema:
type: string
format: uuid
PageParam:
name: page
in: query
required: false
description: >-
The page number for pagination.
schema:
type: integer
minimum: 0
default: 0
PageSizeParam:
name: page_size
in: query
required: false
description: >-
The number of items per page.
schema:
type: integer
minimum: 1
maximum: 100
default: 20
schemas:
CreateFineTuningJobRequest:
type: object
required:
- model
- training_files
properties:
model:
type: string
description: >-
The name of the model to fine-tune.
example: open-mistral-7b
training_files:
type: array
description: >-
A collection of training file IDs with optional weights.
items:
$ref: '#/components/schemas/TrainingFile'
validation_files:
type: array
description: >-
A collection of validation file IDs.
items:
type: string
format: uuid
hyperparameters:
$ref: '#/components/schemas/Hyperparameters'
suffix:
type: string
description: >-
A string that will be added to the fine-tuned model name.
maxLength: 18
auto_start:
type: boolean
description: >-
Whether to automatically start the fine-tuning job after
validation. If false, the job must be started manually.
default: false
integrations:
type: array
description: >-
A list of integrations to enable for the fine-tuning job,
such as Weights and Biases for experiment tracking.
items:
$ref: '#/components/schemas/Integration'
TrainingFile:
type: object
required:
- file_id
properties:
file_id:
type: string
format: uuid
description: >-
The ID of the uploaded training file.
weight:
type: number
description: >-
The weight of this training file in the training mix.
default: 1.0
Hyperparameters:
type: object
properties:
training_steps:
type: integer
description: >-
The number of training steps for the fine-tuning job.
minimum: 1
learning_rate:
type: number
description: >-
The learning rate for the fine-tuning job.
minimum: 0.0
weight_decay:
type: number
description: >-
Weight decay applied during training.
minimum: 0.0
warmup_fraction:
type: number
description: >-
The fraction of training steps to use for warmup.
minimum: 0.0
maximum: 1.0
Integration:
type: object
required:
- type
properties:
type:
type: string
description: >-
The type of integration.
enum:
- wandb
project:
type: string
description: >-
The project name for the integration.
name:
type: string
description: >-
The run name for the integration.
api_key:
type: string
description: >-
The API key for the integration service.
FineTuningJob:
type: object
properties:
id:
type: string
format: uuid
description: >-
The unique identifier for the fine-tuning job.
auto_start:
type: boolean
description: >-
Whether the job was configured to auto-start.
hyperparameters:
$ref: '#/components/schemas/Hyperparameters'
model:
type: string
description: >-
The base model being fine-tuned.
status:
type: string
description: >-
The current status of the fine-tuning job.
enum:
- QUEUED
- STARTED
- VALIDATING
- VALIDATED
- RUNNING
- FAILED_VALIDATION
- FAILED
- SUCCESS
- CANCELLED
- CANCELLATION_REQUESTED
job_type:
type: string
description: >-
The type of fine-tuning job.
created_at:
type: integer
description: >-
Unix timestamp when the job was created.
modified_at:
type: integer
description: >-
Unix timestamp when the job was last modified.
training_files:
type: array
description: >-
The training files used for this job.
items:
type: string
format: uuid
validation_files:
type: array
description: >-
The validation files used for this job.
items:
type: string
format: uuid
fine_tuned_model:
type: string
description: >-
The ID of the fine-tuned model, available after successful
completion.
integrations:
type: array
description: >-
The integrations enabled for this job.
items:
$ref: '#/components/schemas/Integration'
FineTuningJobList:
type: object
properties:
data:
type: array
description: >-
A list of fine-tuning jobs.
items:
$ref: '#/components/schemas/FineTuningJob'
total:
type: integer
description: >-
The total number of fine-tuning jobs.
Error:
type: object
properties:
message:
type: string
description: >-
A human-readable error message.
type:
type: string
description: >-
The type of error.
code:
type: integer
description: >-
The HTTP status code.