Mistral AI Batch API
Batch inference API for processing up to one million requests asynchronously at reduced cost, supporting chat completions, embeddings, FIM, moderations, OCR, classifications, and audio transcriptions.
OpenAPI Specification
openapi: 3.1.0
info:
title: Mistral AI Batch API
description: >-
Batch inference API for processing up to one million requests asynchronously
at reduced cost. Supports chat completions, embeddings, FIM, moderations,
OCR, classifications, and audio transcriptions in batch mode.
version: '1.0'
contact:
name: Mistral AI Support
url: https://docs.mistral.ai/
email: support@mistral.ai
termsOfService: https://mistral.ai/terms/
externalDocs:
description: Mistral AI Batch API Documentation
url: https://docs.mistral.ai/api/endpoint/batch
servers:
- url: https://api.mistral.ai/v1
description: Mistral AI Production
tags:
- name: Batch Jobs
description: Batch inference job operations
security:
- bearerAuth: []
paths:
/batch/jobs:
get:
operationId: listBatchJobs
summary: Mistral AI List batch jobs
description: >-
Retrieve a list of batch inference jobs for the authenticated account.
tags:
- Batch Jobs
parameters:
- name: page
in: query
schema:
type: integer
default: 0
description: Page number for pagination
- name: page_size
in: query
schema:
type: integer
default: 20
description: Number of items per page
- name: status
in: query
schema:
type: string
enum:
- QUEUED
- RUNNING
- SUCCESS
- FAILED
- TIMEOUT_EXCEEDED
- CANCELLATION_REQUESTED
- CANCELLED
description: Filter by job status
responses:
'200':
description: List of batch jobs
content:
application/json:
schema:
$ref: '#/components/schemas/BatchJobList'
'401':
description: Unauthorized
post:
operationId: createBatchJob
summary: Mistral AI Create a batch job
description: >-
Create a new batch inference job. The input file must be a JSONL file
with one request per line, uploaded via the Files API with purpose
set to batch.
tags:
- Batch Jobs
requestBody:
required: true
content:
application/json:
schema:
$ref: '#/components/schemas/CreateBatchJobRequest'
responses:
'200':
description: Batch job created
content:
application/json:
schema:
$ref: '#/components/schemas/BatchJob'
'400':
description: Bad request
'401':
description: Unauthorized
/batch/jobs/{job_id}:
get:
operationId: getBatchJob
summary: Mistral AI Get batch job details
description: Retrieve the current status and details of a batch job.
tags:
- Batch Jobs
parameters:
- $ref: '#/components/parameters/jobId'
responses:
'200':
description: Batch job details
content:
application/json:
schema:
$ref: '#/components/schemas/BatchJob'
'401':
description: Unauthorized
'404':
description: Job not found
/batch/jobs/{job_id}/cancel:
post:
operationId: cancelBatchJob
summary: Mistral AI Cancel a batch job
description: Request cancellation of a running or queued batch job.
tags:
- Batch Jobs
parameters:
- $ref: '#/components/parameters/jobId'
responses:
'200':
description: Batch job cancellation requested
content:
application/json:
schema:
$ref: '#/components/schemas/BatchJob'
'401':
description: Unauthorized
'404':
description: Job not found
components:
securitySchemes:
bearerAuth:
type: http
scheme: bearer
description: Mistral AI API key passed as a Bearer token
parameters:
jobId:
name: job_id
in: path
required: true
description: The unique identifier of the batch job
schema:
type: string
format: uuid
schemas:
CreateBatchJobRequest:
type: object
required:
- input_files
- endpoint
- model
properties:
input_files:
type: array
items:
type: string
format: uuid
description: IDs of JSONL input files uploaded via the Files API
endpoint:
type: string
enum:
- /v1/chat/completions
- /v1/embeddings
- /v1/fim/completions
- /v1/moderations
- /v1/chat/moderations
- /v1/ocr
description: The API endpoint to use for batch processing
model:
type: string
description: The model to use for inference
metadata:
type: object
additionalProperties:
type: string
description: Optional metadata key-value pairs
timeout_hours:
type: integer
default: 24
description: Maximum time in hours before the job times out
BatchJob:
type: object
properties:
id:
type: string
format: uuid
description: Unique identifier for the batch job
object:
type: string
enum:
- batch
input_files:
type: array
items:
type: string
format: uuid
description: Input file IDs
endpoint:
type: string
description: The API endpoint used
model:
type: string
description: The model used
output_file:
type:
- string
- 'null'
format: uuid
description: Output file ID when job completes
error_file:
type:
- string
- 'null'
format: uuid
description: Error file ID if there were errors
status:
type: string
enum:
- QUEUED
- RUNNING
- SUCCESS
- FAILED
- TIMEOUT_EXCEEDED
- CANCELLATION_REQUESTED
- CANCELLED
description: Current status of the batch job
created_at:
type: integer
description: Unix timestamp when the job was created
started_at:
type:
- integer
- 'null'
description: Unix timestamp when the job started processing
completed_at:
type:
- integer
- 'null'
description: Unix timestamp when the job completed
metadata:
type: object
additionalProperties:
type: string
total_requests:
type: integer
description: Total number of requests in the batch
completed_requests:
type: integer
description: Number of completed requests
succeeded_requests:
type: integer
description: Number of successful requests
failed_requests:
type: integer
description: Number of failed requests
BatchJobList:
type: object
properties:
data:
type: array
items:
$ref: '#/components/schemas/BatchJob'
total:
type: integer
object:
type: string
enum:
- list