ai.z.openapi.api.batches

Interface BatchesApi

public interface BatchesApi

Batch Processing API for large-scale data operations Enables efficient processing of millions of requests with cost-effective batch execution Provides up to 50% discount compared to individual API calls with 24-hour completion guarantee Supports up to 50,000 requests per batch with maximum 200MB input file size Ideal for evaluation, classification, and embedding tasks that don't require immediate responses

Method Summary

Modifier and Type	Method and Description
io.reactivex.rxjava3.core.Single<Batch>	batchesCancel(String batchId) Cancel a running batch job Stops the batch processing and marks the job as cancelled (may take up to 10 minutes) Only works for batches in 'validating' or 'in_progress' status
io.reactivex.rxjava3.core.Single<Batch>	batchesCreate(BatchCreateParams batchCreateParams) Create a new batch processing job Submits a large number of requests for efficient batch processing with JSONL format Each request must include a unique custom_id for result mapping Supports /v1/chat/completions and /v1/embeddings endpoints
io.reactivex.rxjava3.core.Single<BatchPage>	batchesList(String after, Integer limit) List all batch jobs with pagination Retrieves a paginated list of all batch processing jobs with filtering options Useful for monitoring multiple batch operations and their completion status
io.reactivex.rxjava3.core.Single<Batch>	batchesRetrieve(String batchId) Retrieve details of a specific batch job Gets comprehensive information about batch processing status and results Status can be: validating, failed, in_progress, finalizing, completed, expired, cancelling, cancelled

Method Detail

batchesCreate

@POST(value="batches")
io.reactivex.rxjava3.core.Single<Batch> batchesCreate(@Body
                                                                             BatchCreateParams batchCreateParams)

Create a new batch processing job Submits a large number of requests for efficient batch processing with JSONL format Each request must include a unique custom_id for result mapping Supports /v1/chat/completions and /v1/embeddings endpoints

Parameters:

batchCreateParams - Batch configuration including input file, endpoint, and processing options

Returns:

Batch job information with ID, status, and processing details

batchesRetrieve

@GET(value="batches/{batch_id}")
io.reactivex.rxjava3.core.Single<Batch> batchesRetrieve(@Path(value="batch_id")
                                                                                         String batchId)

Retrieve details of a specific batch job Gets comprehensive information about batch processing status and results Status can be: validating, failed, in_progress, finalizing, completed, expired, cancelling, cancelled

Parameters:

batchId - Unique identifier of the batch job to retrieve

Returns:

Batch job details including progress, completion status, output_file_id, and error_file_id

batchesList

@GET(value="batches")
io.reactivex.rxjava3.core.Single<BatchPage> batchesList(@Query(value="after")
                                                                              String after,
                                                                              @Query(value="limit")
                                                                              Integer limit)

List all batch jobs with pagination Retrieves a paginated list of all batch processing jobs with filtering options Useful for monitoring multiple batch operations and their completion status

Parameters:

after - Cursor for pagination to get batches after this point

limit - Maximum number of batch jobs to return per page (default: 20)

Returns:

Paginated list of batch jobs with status, request counts, and metadata

batchesCancel

@POST(value="batches/{batch_id}/cancel")
io.reactivex.rxjava3.core.Single<Batch> batchesCancel(@Path(value="batch_id")
                                                                                               String batchId)

Cancel a running batch job Stops the batch processing and marks the job as cancelled (may take up to 10 minutes) Only works for batches in 'validating' or 'in_progress' status

Parameters:

batchId - Unique identifier of the batch job to cancel

Returns:

Updated batch job status after cancellation with final request counts