Serverless Containers

Get a container deployment template

get/v1/container-deployment-templates/{template_id}

Request

  • Base URL: https://api.verda.com
  • URL: https://api.verda.com/v1/container-deployment-templates/{template_id}
  • Auth: HTTP bearer

Path parameters

template_idstring required
Example:qwen3-5-9b

Id of the template

Response

The template with its variants and features

idstring required

Stable, URL-safe template identifier used in template paths

namestring required

Human-readable name of the model

descriptionstring required

Description of the model

providerstring required

Provider of the model

enginestring required

Serving engine used by the template

model_repositorystring required

Hugging Face repository the default variant loads

parametersstring required

Parameter count of the model (display string)

context_lengthnumber required

Maximum context length in tokens

input_modalitiesstring[] required

Accepted input modalities

output_modalitiesstring[] required

Produced output modalities

tasksstring[] required

Coarse task labels

requires_hugging_face_tokenboolean required

Whether a Hugging Face token is required to deploy the model (gated checkpoint)

min_vram_gbnumber required

Minimum total GPU memory in GB for the default variant

revisionobject nullable required

Pinned Hugging Face revision of the default repository. Null tracks the main branch

Example response

{
  "id": "qwen3-5-9b",
  "name": "Qwen3.5 9B",
  "description": "Dense multimodal model with gated delta networks hybrid attention and 262K context. Single-GPU serving.",
  "provider": "Qwen",
  "engine": "vllm",
  "model_repository": "Qwen/Qwen3.5-9B",
  "parameters": "9B",
  "context_length": 262144,
  "input_modalities": [
    "text",
    "image"
  ],
  "output_modalities": [
    "text"
  ],
  "tasks": [
    "text",
    "multimodal"
  ],
  "min_vram_gb": 22,
  "recommended_compute": {
    "name": "H100",
    "size": 1
  },
  "variants": [
    {
      "id": "fp8",
      "precision": "fp8",
      "model_repository": "Qwen/Qwen3.5-9B-FP8",
      "min_vram_gb": 12,
      "is_default": true,
      "description": "FP8 weights and KV cache, fits a single 16 GB GPU",
      "recommended_compute": {
        "name": "H100",
        "size": 1
      }
    }
  ],
  "features": [
    {
      "id": "tool_calling",
      "name": "Tool calling",
      "description": "Enables tool calling with the Qwen chat template",
      "is_default": true,
      "conflicts_with": []
    }
  ]
}

Changes

No recorded changes to this operation across all 1 revision of this API.