{
  "collected_at_local": "2026-10-03T18:09:34.5489107+03:00",
  "source": "http://127.0.0.1:11434/api/ps",
  "loaded_models": [
    {
      "name": "qwen2.5:1.5b-instruct-q4_K_M",
      "model": "qwen2.5:1.5b-instruct-q4_K_M",
      "loaded_size_bytes": 1169980128,
      "vram_size_bytes": 0,
      "parameter_size": "1.5B",
      "quantization": "Q4_K_M",
      "context_length": 4096,
      "expires_at": "2026-10-03T18:14:33.40604+03:00"
    },
    {
      "name": "gemma4:e2b",
      "model": "gemma4:e2b",
      "loaded_size_bytes": 6733158152,
      "vram_size_bytes": 0,
      "parameter_size": "5.1B",
      "quantization": "Q4_K_M",
      "context_length": 4096,
      "expires_at": "2026-10-03T18:12:14.9381872+03:00"
    }
  ],
  "llama_server_processes": [
    {
      "process_id": 41288,
      "working_set_bytes": 3017682944,
      "private_bytes": 3743903744
    },
    {
      "process_id": 55716,
      "working_set_bytes": 1193771008,
      "private_bytes": 1283440640
    }
  ],
  "note": "Point-in-time snapshot, not peak memory. Process-to-model mapping and total host RAM are not asserted."
}