• Original Model Card
• vLLM Documentation

⚠️ Warning: ibm-granite/granite-4.1-30b-base is a base model hence it doesn't have a pre-defined chat template as it's not fine-tuned for chat applications but rahter just pre-trained, meaning that it's likely not optimal for consumer facing inference or downstream applications.

Completions API

⚠️ Warning: note that the OpenAI Completions API is a legacy API, despite it being supported in vLLM; but it's the only way to get completions for base models as those don't have a chat template.

Send Request

curl <AZUREML_ENDPOINT_URL> \
    -X POST \
    -H "Authorization: Bearer <AZUREML_TOKEN>" \
    -H "Content-Type: application/json" \
    -d '{"model":"ibm-granite/granite-4.1-30b-base","prompt":"What is Deep Learning?"}'

Supported Parameters

• model (string): Model ID used to generate the response, in this case since only a single model is deployed within the same endpoint you can either set it to ibm-granite/granite-4.1-30b-base or leave it blank instead.
• prompt (string or array): The prompt(s) to generate completions for, encoded as a string, array of strings, array of tokens, or array of token arrays. Note that <|endoftext|> is the document separator that the model sees during training, so if a prompt is not specified the model will generate as if from the beginning of a new document.

Example payload

{
  "model": "ibm-granite/granite-4.1-30b-base",
  "prompt": "What is Deep Learning?",
  "max_tokens": 256,
  "temperature": 0.6
}