Developers / Docs / API reference / Models and status

Models and status

List the models a deployment serves, in OpenAI and Anthropic format, and check that the service is up. For examples and explanations, see the Models guide.

List models

GET /api/v1/models

The models that are online now. No API key needed. Add ?type= to filter by ark_type.

Model object

idstring

The id to use in requests.

object, created, owned_by, shutdown_date

"model", a fixed timestamp, "ark" and null.

ark_typestringARK extension

causal_lm, multimodal, embedding or speech_transcription.

max_tokensintegerARK extension

Context length: the most tokens per request or stateful session.

ark_reasoningbooleanARK extension

Whether the model can reason. Omitted briefly after a restart.

GET /api/v1/models

{
  "object": "list",
  "data": [{
    "id": "arklabsai/Qwen3.5-122B-A10B",
    "object": "model",
    "created": 1754035200,
    "owned_by": "ark",
    "shutdown_date": null,
    "ark_type": "multimodal",
    "max_tokens": 262144,
    "ark_reasoning": true
  }, …]
}

Retrieve model

GET /api/v1/models/{model_id}

One model object. The id contains a slash and is used as is. An unknown id returns 404 model_not_found.

Request

curl https://api.ark-labs.cloud/api/v1/models/arklabsai/bge-m3

List models (Anthropic format)

GET /api/anthropic/v1/models

All online models in Anthropic's list format. No API key needed. Page with limit (1–1000, default 20), after_id and before_id. GET /api/anthropic/v1/models/{model_id} returns one model.

Response

{
  "data": [{
    "type": "model",
    "id": "arklabsai/Qwen3.6-35B-A3B",
    "display_name": "arklabsai/Qwen3.6-35B-A3B",
    "created_at": "2025-08-01T10:00:00Z"
  }, …],
  "has_more": false,
  "first_id": "arklabsai/Qwen3.6-35B-A3B",
  "last_id": "arklabsai/whisper-large-v3-turbo"
}

Status

GET /api/v1/status

A health check for monitoring. No API key, empty body: 200 while at least one model is fully online, 500 otherwise.

Request

curl -i https://api.ark-labs.cloud/api/v1/status