Models API

List and monitor AI models in the Spice runtime

Overview

The Models API provides an endpoint to list all machine learning and language models available in the runtime, compatible with the OpenAI models API format.

List Models

GET /v1/models

Returns a list of all models (both ML and LLM) available in the runtime.

Query Parameters

status · boolean · default: false

Include the current status of each model. Possible values:

• ready - Model is ready for inference
• loading - Model is being loaded
• error - Model encountered an error

format · string · default: "json"

Response format: json or csv

metadata_fields · string

Comma-separated list of metadata fields to include. Available fields:

• supports_responses_api - Whether the model supports the responses API

Response

object · string

Always "list"

data · array<object>

Array of model information objects.

id · string

The model identifier (e.g., gpt-4, text-embedding-ada-002)

object · string

Always "model"

owned_by · string

The provider or owner of the model (e.g., openai, openai-internal, spiceai)

datasets · array<string>

List of datasets associated with this model (for embedding models). null if no datasets.

status · string

Current status of the model (only when status=true query parameter is set)

error · object

Error information when status is error (only when status=true)

category · string

Error category (e.g., model, worker)

type · string

Error type (e.g., auth, connection, loading)

code · string

Stable error code (e.g., model.auth, model.loading)

error_message · string

Human-readable error message (only when status=true and status is error)

metadata · object

Additional model metadata (only when metadata_fields query parameter is set)

supports_responses_api · boolean

Whether the model supports the responses API

Response Example (JSON)

{
  "object": "list",
  "data": [
    {
      "id": "gpt-4",
      "object": "model",
      "owned_by": "openai",
      "datasets": null,
      "status": "ready",
      "error": null,
      "error_message": null
    },
    {
      "id": "text-embedding-ada-002",
      "object": "model",
      "owned_by": "openai-internal",
      "datasets": ["text-dataset-1", "text-dataset-2"],
      "status": "error",
      "error": {
        "category": "model",
        "type": "auth",
        "code": "model.auth"
      },
      "error_message": "Invalid API key"
    }
  ]
}

Response Example (CSV)

id,object,owned_by,datasets,status,error,error_message
gpt-4,model,openai,,ready,,
text-embedding-ada-002,model,openai-internal,"text-dataset-1,text-dataset-2",error,model.auth,Invalid API key

Response with Metadata

When requesting metadata_fields=supports_responses_api:

{
  "object": "list",
  "data": [
    {
      "id": "gpt-4",
      "object": "model",
      "owned_by": "openai",
      "datasets": null,
      "status": "ready",
      "error": null,
      "error_message": null,
      "metadata": {
        "supports_responses_api": true
      }
    },
    {
      "id": "text-embedding-ada-002",
      "object": "model",
      "owned_by": "openai-internal",
      "datasets": ["text-dataset-1", "text-dataset-2"],
      "status": "ready",
      "error": null,
      "error_message": null,
      "metadata": {
        "supports_responses_api": false
      }
    }
  ]
}

Status Codes

• 200 OK - Models retrieved successfully
• 500 Internal Server Error - App not initialized or unexpected error

Error Response (500)

App not initialized

Examples

List All Models

curl http://localhost:8090/v1/models

List Models with Status

curl http://localhost:8090/v1/models?status=true

List Models with Metadata

curl http://localhost:8090/v1/models?metadata_fields=supports_responses_api

Get CSV Format

curl http://localhost:8090/v1/models?format=csv

Combined Parameters

curl http://localhost:8090/v1/models?status=true&format=json&metadata_fields=supports_responses_api

OpenAI Compatibility

This endpoint follows the OpenAI models API format, making it compatible with OpenAI client libraries:

import openai

client = openai.OpenAI(
    api_key="your-key",
    base_url="http://localhost:8090/v1"
)

# List all models
models = client.models.list()
for model in models.data:
    print(f"{model.id} - {model.owned_by}")

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'your-key',
  baseURL: 'http://localhost:8090/v1',
});

// List all models
const models = await client.models.list();
models.data.forEach(model => {
  console.log(`${model.id} - ${model.owned_by}`);
});