Translate text

For translation tasks, Generative AI on Vertex AI offers two specialized translation models from the Cloud Translation API:

• Translation LLM - Google's newest highest quality LLM-style translation offering. Achieves the highest quality translation while serving at reasonable latencies (~3x faster than Gemini 2.0 Flash).

• Cloud Translation Neural Machine Translation (NMT) model - Google's premier real-time translation offering achieving ~100ms latency translations. It achieves the highest quality of all models benchmarked serving at comparable latencies and continues to see ongoing quality advancements. NMT can achieve latencies up to 20x faster than Gemini 2.0 Flash.

Key advantages & differentiators of the Translation LLM

• Unmatched Translation Quality - Translation LLM offers the highest translation quality achieving significantly higher performance on benchmarks compared to other benchmarked models. Translation LLM is far more likely to significantly rewrite a sentence to make it sound more natural in the target language, rather than giving less natural "word-for-word" translations that are often seen in other translation models.
• Superior Quality/Latency Trade off - Translation LLM provides LLM-powered translations at latencies significantly better than Gemini 2.0 Flash. Although Translation LLM has higher latencies than the NMT model, it typically provides higher quality responses for a broad range of applications.

Model feature comparison

Usage

This section shows you how to use Vertex AI Studio to rapidly translate text from one language to another. You can use the Translation LLM or the NMT model to translate text by using the Google Cloud console or API. Note that the languages that each model supports can vary. Before you request translations, check that the model you're using supports your source and target languages.

Console

1. In the Vertex AI section of the Google Cloud console, go to the Translate text page in Vertex AI Studio.

   Go to Vertex AI Studio

2. In the Run settings pane, select a translation model in the Model field.

3. To change the model settings (such as temperature), expand Advanced.

4. Set the source and target languages.

5. In the input field, enter the text to translate.

6. Click Submit.

7. To get the code or curl command that demonstrate how to request translations, click code Get code.

Note that in Vertex AI Studio, the Translation LLM lets you provide example translations to tailor model responses to more closely match your style, tone, and industry domain. The model uses your examples as few-shot context before translating your text.

API

Select the model to use for your translations.

POST https://LOCATION-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/LOCATION/publishers/google/models/cloud-translate-text:predict

{
  "instances": [
    {
      "source_language_code": "SOURCE_LANGUAGE_CODE",
      "target_language_code": "TARGET_LANGUAGE_CODE",
      "contents": [
        "SOURCE_TEXT"
      ],
      "mimeType": "MIME_TYPE",
      "model": "projects/PROJECT_ID/locations/LOCATION/models/general/translation-llm"
    }
  ]
}

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "x-goog-user-project: PROJECT_NUMBER_OR_ID" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://LOCATION-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/LOCATION/publishers/google/models/cloud-translate-text:predict"

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred"; "x-goog-user-project" = "PROJECT_NUMBER_OR_ID" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://LOCATION-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/LOCATION/publishers/google/models/cloud-translate-text:predict" | Select-Object -Expand Content

{
  "predictions": [
    {
      "translations": [
        {
          "translatedText": "TRANSLATED_TEXT",
          "model": "projects/PROJECT_ID/locations/LOCATION/models/general/translation-llm"
        }
      ]
    }
  ]
}

async function translate() {
  const request = {
    instances: [{
      source_language_code: SOURCE_LANGUAGE_CODE,
      target_language_code: TARGET_LANGUAGE_CODE,
      contents: [SOURCE_TEXT],
      model: "projects/PROJECT_ID/locations/LOCATION/models/general/translation-llm"
    }]
  };
  const {google} = require('googleapis');
  const aiplatform = google.cloud('aiplatform');
  const endpoint = aiplatform.predictionEndpoint('projects/PROJECT_ID/locations/LOCATION/publishers/google/models/cloud-translate-text');

  const [response] = await endpoint.predict(request)
  console.log('Translating')
  console.log(response)
}
      

from google.cloud import aiplatform

def translate():
  # Create a client
  client_options = {"api_endpoint": "LOCATION-aiplatform.googleapis.com"}
  client = aiplatform.gapic.PredictionServiceClient(client_options=client_options)
  
  # Initialize the request
  endpoint_id = f"projects/PROJECT_ID/locations/LOCATION/publishers/google/models/cloud-translate-text"
  instances=[{
    "model": "projects/PROJECT_ID/locations/LOCATION/models/general/translation-llm",
    "source_language_code": 'SOURCE_LANGUAGE_CODE',
    "target_language_code": 'TARGET_LANGUAGE_CODE',
    "contents": ["SOURCE_TEXT"],
  }]

  # Make the request
  response = client.predict(instances=instances, endpoint=endpoint_id)
  # Handle the response
  print(response)
      

POST https://translation.googleapis.com/v3/projects/PROJECT_ID:translateText

{
  "sourceLanguageCode": "SOURCE_LANGUAGE",
  "targetLanguageCode": "TARGET_LANGUAGE",
  "contents": ["SOURCE_TEXT1", "SOURCE_TEXT2"]
}

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "x-goog-user-project: PROJECT_NUMBER_OR_ID" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://translation.googleapis.com/v3/projects/PROJECT_ID:translateText"

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred"; "x-goog-user-project" = "PROJECT_NUMBER_OR_ID" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://translation.googleapis.com/v3/projects/PROJECT_ID:translateText" | Select-Object -Expand Content

{
  "translations": [
    {
      "translatedText": "TRANSLATED_TEXT1"
    },
    {
      "translatedText": "TRANSLATED_TEXT2"
    }
  ]
}

/**
 * TODO(developer): Uncomment these variables before running the sample
 */
// const projectId = 'YOUR_PROJECT_ID';
// const location = 'global';
// const text = 'text to translate';

// Imports the Google Cloud Translation library
const {TranslationServiceClient} = require('@google-cloud/translate');

// Instantiates a client
const translationClient = new TranslationServiceClient();

async function translateText() {
  // MIME type of the content to translate
  // Supported MIME types:
  // https://cloud.google.com/translate/docs/supported-formats
  const mimeType = 'text/plain';

  // Construct request
  const request = {
    parent: `projects/${projectId}/locations/${location}`,
    contents: [text],
    mimeType: mimeType,
    sourceLanguageCode: 'en',
    targetLanguageCode: 'sr-Latn',
  };

  // Run request
  const [response] = await translationClient.translateText(request);

  for (const translation of response.translations) {
    console.log(`Translation: ${translation.translatedText}`);
  }
}

translateText();

import os

# Import the Google Cloud Translation library.
from google.cloud import translate_v3

PROJECT_ID = os.environ.get("GOOGLE_CLOUD_PROJECT")


def translate_text(
    text: str = "YOUR_TEXT_TO_TRANSLATE",
    source_language_code: str = "en-US",
    target_language_code: str = "fr",
) -> translate_v3.TranslationServiceClient:
    """Translate Text from a Source language to a Target language.
    Args:
        text: The content to translate.
        source_language_code: The code of the source language.
        target_language_code: The code of the target language.
            For example: "fr" for French, "es" for Spanish, etc.
            Find available languages and codes here:
            https://cloud.google.com/translate/docs/languages#neural_machine_translation_model
    """

    # Initialize Translation client.
    client = translate_v3.TranslationServiceClient()
    parent = f"projects/{PROJECT_ID}/locations/global"

    # MIME type of the content to translate.
    # Supported MIME types:
    # https://cloud.google.com/translate/docs/supported-formats
    mime_type = "text/plain"

    # Translate text from the source to the target language.
    response = client.translate_text(
        contents=[text],
        parent=parent,
        mime_type=mime_type,
        source_language_code=source_language_code,
        target_language_code=target_language_code,
    )

    # Display the translation for the text.
    # For example, for "Hello! How are you doing today?":
    # Translated text: Bonjour comment vas-tu aujourd'hui?
    for translation in response.translations:
        print(f"Translated text: {translation.translated_text}")

    return response

Custom translations

Customize responses from the Translation LLM by providing your own example translations. Custom translations only work with the Translation LLM.

You can request customized translation through the Vertex AI Studio console or API with one difference. The console supports custom translations only when you provide examples in a TMX or TSV file. The API supports custom translations only when you provide examples (up to 5 sentence pairs) inline as part of the translation request.

Data requirements

If you provide example translations in a file for the Google Cloud console, the examples must be written as segment pairs in a TMX or TSV file. Each pair includes a source language segment and its translated counterpart. For more information, see Prepare example translations in the Cloud Translation documentation.

To get the most accurate results, include specific examples from a wide variety of scenarios. You must include at least five sentence pairs but no more than 10,000 pairs. Also, a segment pair can be at most a total of 512 characters.

Console

1. In the Vertex AI section of the Google Cloud console, go to the Translate text page in Vertex AI Studio.

   Go to Vertex AI Studio

2. In the Run settings pane, configure your translation settings.

   1. In the Model field, select Translation LLM.
   2. To change the temperature, expand Advanced.

3. Click Add examples.

   1. Select a local file or a file from Cloud Storage. Vertex AI Studio determines the source and target languages from your file.
   2. Select the number of examples for the model to use before generating a response.

   The number of examples you select counts toward the input character limit per request of 3,000.

4. In the input field, enter the text to translate.

5. Click Submit.

   Vertex AI automatically selects your specified number of reference sentences that are most similar to your input. The translation model identifies patterns from your examples and then applies those patterns when generating a response.

   The output limit per request is 3,000 characters. Any text beyond this limit is dropped.

6. To get the code or curl command that demonstrate how to request translations, click code Get code.

API

To request custom translations, include up to five reference sentence pairs in your translation request. The translation model uses all of them to identify patterns from your examples and then applies those patterns when generating a response.

POST https://LOCATION-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/LOCATION/publishers/google/models/translate-llm:predict

{
  "instances": [
    {
      "reference_sentence_config": {
        "reference_sentence_pair_lists": [
          {
            "reference_sentence_pairs": [
              {
                "source_sentence": "REFERENCE_SOURCE_1_1",
                "target_sentence": "REFERENCE_TARGET_1_1"
              },
              {
                "source_sentence": "REFERENCE_SOURCE_1_2",
                "target_sentence": "REFERENCE_SOURCE_1_2"
              }
            ]
          }
        ],
        "source_language_code": "SOURCE_LANGUAGE_CODE",
        "target_language_code": "TARGET_LANGUAGE_CODE"
      },
      "content": [
        "SOURCE_TEXT"
      ],
      "mimeType": "MIME_TYPE"
    }
  ]
}

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "x-goog-user-project: PROJECT_NUMBER_OR_ID" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://LOCATION-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/LOCATION/publishers/google/models/translate-llm:predict"

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred"; "x-goog-user-project" = "PROJECT_NUMBER_OR_ID" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://LOCATION-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/LOCATION/publishers/google/models/translate-llm:predict" | Select-Object -Expand Content

{
  "predictions": [
    {
      "languageCode": "TARGET_LANGUAGE",
      "translations": [
        {
          "translatedText": "TRANSLATED_TEXT"
        }
      ]
    }
  ]
}

async function translate() {
  const request = {
    instances: [{
        "reference_sentence_config": {
          "reference_sentence_pair_lists": [{
            "reference_sentence_pairs": [{
              "source_sentence": 'SAMPLE_REFERENCE_SOURCE_1',
              "target_sentence": 'SAMPLE_REFERENCE_TARGET_1'
            },
            "reference_sentence_pairs": {
              "source_sentence": 'SAMPLE_REFERENCE_SOURCE_2',
              "target_sentence": 'SAMPLE_REFERENCE_TARGET_2'
            }]
          }],
          "source_language_code": 'SOURCE_LANGUAGE_CODE',
          "target_language_code": 'TARGET_LANGUAGE_CODE'
        },
        "contents": ["SOURCE_TEXT"]
    }]
  };
  const {google} = require('googleapis');
  const aiplatform = google.cloud('aiplatform');
  const endpoint = aiplatform.predictionEndpoint('projects/PROJECT_ID/locations/LOCATION/publishers/google/models/translate-llm');

  const [response] = await endpoint.predict(request)
  console.log('Translating')
  console.log(response)
}
  

from google.cloud import aiplatform
from google.protobuf.json_format import MessageToDict

def translate():
  # Create a client
  client_options = {"api_endpoint": "LOCATION-aiplatform.googleapis.com"}
  client = aiplatform.gapic.PredictionServiceClient(client_options=client_options)

  # Initialize the request
  endpoint_id = f"projects/PROJECT_ID/locations/LOCATION/publishers/google/models/translate-llm"
  instances=[{
      "reference_sentence_config": {
        "reference_sentence_pair_lists": [{
          "reference_sentence_pairs": [{
            "source_sentence": 'SAMPLE_REFERENCE_SOURCE_1',
            "target_sentence": 'SAMPLE_REFERENCE_TARGET_1'
          },
          {
            "source_sentence": 'SAMPLE_REFERENCE_SOURCE_2',
            "target_sentence": 'SAMPLE_REFERENCE_TARGET_2'
          }]
        }],
        "source_language_code": 'SOURCE_LANGUAGE_CODE',
        "target_language_code": 'TARGET_LANGUAGE_CODE'
      },
      "content": ["SOURCE_TEXT"]
  }]
  # Make the request
  response = client.predict(
      endpoint=endpoint_id, instances=instances,
  )
  # Handle the response
  print(response)

  # The predictions are a google.protobuf.Value representation of the model's predictions.
  predictions = MessageToDict(response._pb)['predictions']
  for prediction in predictions:
      print(prediction['translations'])
  

You can also use the Cloud Translation API to create a dataset and import your example sentence pairs. When you use the Cloud Translation API to request translations, you can include your dataset to customize responses. The dataset persists and can be reused with multiple translation requests. For more information, see Request adaptive translations in the Cloud Translation documentation.

Supported languages

Translation LLM

With the Translation LLM, you can translate to and from any of the following languages.

NMT

For information about which languages the the Cloud Translation NMT model support, refer to the following documentation:

• Translation NMT model language support