> ## Documentation Index
> Fetch the complete documentation index at: https://docs.portkey.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Batches

> Perform batch inference with Vertex AI

With Prisma AIRS AI Gateway, you can perform batch inference operations with Vertex AI models. This is the most efficient way to:

* Process large volumes of data with Vertex AI models
* Test your data with different foundation models
* Perform A/B testing with different foundation models

## Before You Start

1. **AI Gateway API key** and a **Vertex AI provider** configured in Model Catalog.
2. A **GCS bucket** in the same region as your model + `aiplatform-service-agent` permission on the file.
3. *(Only for gateway-native batching)* A **AI Gateway File** (`input_file_id`).
4. Familiarity with the [Create Batch OpenAPI spec](/docs/api-reference/inference-api/batch/create-batch).

<Info>
  The AI Gateway supports **two modes** on Vertex:

  * **[Provider Batch API](#using-vertex-batch-api-through-the-ai-gateway)** (cheapest, completion\_window: `24h`, `48h`, etc.)
  * **[AI Gateway Batch API](/docs/aigw/product/ai-gateway/batches#portkey-batch-api-mode)** (fast, provider-agnostic, completion\_window: `immediate`)
</Info>

## Using Vertex Batch API through the AI Gateway

### Upload a file for batch inference

<CodeGroup>
  ```bash curl theme={"system"}
  curl -X POST --header 'Authorization: Bearer <portkey_api_key>' \
   --header 'x-portkey-provider: @your-vertex-provider' \
   --header 'x-portkey-vertex-storage-bucket-name: <bucket_name>' \
   --header 'x-portkey-provider-file-name: <file_name>.jsonl' \
   --header 'x-portkey-provider-model: <model_name>' \
   --form 'purpose="batch"' \
   --form 'file=@dataset.jsonl' \
   'https://aigw.portkey.ai/v1/files'
  ```

  ```python OpenAI Python theme={"system"}
  from openai import OpenAI

  openai = OpenAI(
      api_key="PORTKEY_API_KEY",
      base_url="https://aigw.portkey.ai/v1",
      default_headers={"x-portkey-provider": "@VERTEX_PROVIDER", "x-portkey-vertex-storage-bucket-name": "your_bucket_name", "x-portkey-provider-file-name": "your_file_name.jsonl", "x-portkey-provider-model": "gemini-1.5-flash-001"}
  )

  # Upload a file for batch inference
  file = openai.files.create(
      file=open("dataset.jsonl", "rb"),
      purpose="batch"
  )

  print(file)
  ```

  ```javascript OpenAI NodeJS theme={"system"}
  import OpenAI from 'openai';
  import * as fs from 'fs';

  const openai = new OpenAI({
      apiKey: "PORTKEY_API_KEY",
      baseURL: "https://aigw.portkey.ai/v1",
      defaultHeaders: { "x-portkey-provider": "@VERTEX_PROVIDER", "x-portkey-vertex-storage-bucket-name": "your_bucket_name", "x-portkey-provider-file-name": "your_file_name.jsonl", "x-portkey-provider-model": "gemini-1.5-flash-001" }
  });

  (async () => {
      // Upload a file for batch inference
      const file = await openai.files.create({
          file: fs.createReadStream("dataset.jsonl"),
          purpose: "batch"
      });

      console.log(file);
  })();
  ```
</CodeGroup>

### Create a batch job

<CodeGroup>
  ```bash curl theme={"system"}
  curl -X POST --header 'Content-Type: application/json' \
   --header 'Authorization: Bearer <portkey_api_key>' \
   --data \
   $'{"input_file_id": "<file_id>", "endpoint": "/v1/chat/completions", "completion_window": "24h", "model":"@your-vertex-provider/gemini-1.5-flash-001"}' \
  'https://aigw.portkey.ai/v1/batches'
  ```

  ```python OpenAI Python theme={"system"}
  from openai import OpenAI

  openai = OpenAI(
      api_key="PORTKEY_API_KEY",
      base_url="https://aigw.portkey.ai/v1"
  )

  # Create a batch inference job
  batch_job = openai.batches.create(
      input_file_id="<file_id>", # File ID from the upload step
      endpoint="/v1/chat/completions", # API endpoint to use
      completion_window="24h", # Time window for completion
      model="@VERTEX_PROVIDER/gemini-1.5-flash-001"
  )

  print(batch_job)
  ```

  ```javascript OpenAI NodeJS theme={"system"}
  import OpenAI from 'openai';

  const openai = new OpenAI({
      apiKey: "PORTKEY_API_KEY",
      baseURL: "https://aigw.portkey.ai/v1"
  });

  (async () => {
      // Create a batch inference job
      const batchJob = await openai.batches.create({
          input_file_id: "<file_id>", // File ID from the upload step
          endpoint: "/v1/chat/completions", // API endpoint to use
          completion_window: "24h", // Time window for completion
          model: "@VERTEX_PROVIDER/gemini-1.5-flash-001"
      });

      console.log(batchJob);
  })();
  ```
</CodeGroup>

### List batch jobs

<CodeGroup>
  ```bash curl theme={"system"}
  curl -X GET --header 'Authorization: Bearer <portkey_api_key>' \
  --header 'x-portkey-provider: @your-vertex-provider' \
  'https://aigw.portkey.ai/v1/batches'
  ```

  ```python OpenAI Python theme={"system"}
  from openai import OpenAI

  openai = OpenAI(
      api_key="PORTKEY_API_KEY",
      base_url="https://aigw.portkey.ai/v1",
      default_headers={"x-portkey-provider": "@VERTEX_PROVIDER"}
  )

  # List all batch jobs
  jobs = openai.batches.list(
      limit=10  # Optional: Number of jobs to retrieve (default: 20)
  )

  print(jobs)
  ```

  ```javascript OpenAI NodeJS theme={"system"}
  import OpenAI from 'openai';

  const openai = new OpenAI({
      apiKey: "PORTKEY_API_KEY",
      baseURL: "https://aigw.portkey.ai/v1",
      defaultHeaders: { "x-portkey-provider": "@VERTEX_PROVIDER" }
  });

  (async () => {
      // List all batch jobs
      const jobs = await openai.batches.list({
          limit: 10  // Optional:
      });

      console.log(jobs);
  })();
  ```
</CodeGroup>

### Get a batch job

<CodeGroup>
  ```bash curl theme={"system"}
  curl -X GET --header 'Authorization: Bearer <portkey_api_key>' \
  --header 'x-portkey-provider: @your-vertex-provider' \
  'https://aigw.portkey.ai/v1/batches/<job_id>'
  ```

  ```python OpenAI Python theme={"system"}
  from openai import OpenAI

  openai = OpenAI(
      api_key="PORTKEY_API_KEY",
      base_url="https://aigw.portkey.ai/v1",
      default_headers={"x-portkey-provider": "@VERTEX_PROVIDER"}
  )

  # Retrieve a specific batch job
  job = openai.batches.retrieve(
      "job_id"  # The ID of the batch job to retrieve
  )

  print(job)
  ```

  ```javascript OpenAI NodeJS theme={"system"}
  import OpenAI from 'openai';

  const openai = new OpenAI({
      apiKey: "PORTKEY_API_KEY",
      baseURL: "https://aigw.portkey.ai/v1",
      defaultHeaders: { "x-portkey-provider": "@VERTEX_PROVIDER" }
  });

  (async () => {
      // Retrieve a specific batch job
      const job = await openai.batches.retrieve(
          "job_id"  // The ID of the batch job to retrieve
      );

      console.log(job);
  })();
  ```
</CodeGroup>

### Get batch job output

<CodeGroup>
  ```bash curl theme={"system"}
  curl -X GET --header 'Authorization: Bearer <portkey_api_key>' \
  --header 'x-portkey-provider: @your-vertex-provider' \
  'https://aigw.portkey.ai/v1/batches/<job_id>/output'
  ```
</CodeGroup>


## Related topics

- [Batches](/docs/aigw/integrations/llms/claude-platform-aws/batches.md)