Skip to content

Listing Models

Use the models endpoint to discover which AI models are currently available on your OptimaGPT deployment.

Endpoint

GET /optima/v1/models

Requires authentication. See Authentication.

Response

The response contains two arrays: one for language models (used with /v1/chat/completions) and one for embedding models (used with /v1/embeddings).

Example response: 

{
  "object": "list",
  "chat_models": [
    {
      "id": "llama-3.1-8b-q4",
      "name": "Llama 3.1 8B Q4",
      "object": "chat_model",
      "owned_by": "optima",
      "ctx_max": 8192,
      "description": "Llama 3.1 8B, Q4_K_M quantisation"
    },
    {
      "id": "qwen2.5-32b-q4",
      "name": "Qwen 2.5 32B Q4",
      "object": "chat_model",
      "owned_by": "optima",
      "ctx_max": 16384,
      "description": null
    }
  ],
  "embed_models": [
    {
      "id": "nomic-embed-text",
      "name": "Nomic Embed Text",
      "object": "embedding_model",
      "owned_by": "optima",
      "dimensions": 768,
      "description": null
    }
  ]
}

Chat model fields

Field Description
id The model identifier to use in API requests
name Human-readable display name
object Always "chat_model"
owned_by Always "optima"
ctx_max Maximum context window size in tokens
description Optional description set in the executor configuration

Embedding model fields

Field Description
id The model identifier to use in API requests
name Human-readable display name
object Always "embedding_model"
owned_by Always "optima"
dimensions The dimensionality of the output embedding vectors
description Optional description set in the executor configuration

Which models appear here

Only models with a running executor appear in the response. Models that are installed but stopped do not appear. If a model you expect to see is missing, check with your administrator that its executor is online.

Using the model ID

The id field from this response is what you pass as the model parameter in /v1/chat/completions and /v1/embeddings requests.