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