Inspect running models

GET /v1/models/running

Reports runtime state rather than catalog presence. During a cold load or a model switch runtime.state can be loading with a temporarily empty model list, which is not completed unloading. Builds containing the queue-status change also return runtime.queues.

Guide and examplesModels and service All endpoints

Responses

200Runtime state. The contents of each model's status are model-specific.

objectstringlist

dataobject[]

idstring

statusobjectModel-specific runtime status.

runtimeobject

statestringrunning, or loading during a cold load or a model switch, when the model list is temporarily empty.

queuesobjectOptional; present only on builds containing the queue-status change. Admission queues are http_transcribe, http_alignment, http_segmentation, http_normalization, http_diarization, http_speaker_embedding, http_broadcast and http_realtime; native_transcribe and native_broadcast describe model ownership across REST, Realtime, MCP and preloading.

{
  "object": "list",
  "data": [
    {
      "id": "Qwen/Qwen3.5-4B",
      "status": {
        "value": "loaded"
      }
    }
  ],
  "runtime": {
    "state": "running"
  }
}

401Invalid or missing API key.

403Host or origin is not allowed, or the license was rejected. Inspect error.code to tell them apart.

503A required local model is still being prepared (model_downloading) or the service is busy (service_busy). Honor Retry-After when present.