{
  "model_id": "m3",
  "model_tag": "phi4:14b",
  "box_class": "cpu-server",
  "started_utc": "2026-08-25T15:11:42Z",
  "nproc": 64,
  "num_predict": 256,
  "prompt_sha256": "90eedd0c53f9554ae3837674504fcb7090013d9a432a653352183c0f25a7ce5c",
  "nvidia_before": "0, 38764 MiB, 0 %",
  "ollama_version": "0.32.15",
  "runs": [
    {
      "ttft_ms": 105.23,
      "first_token_was_thinking": false,
      "wall_s": 24.2159,
      "eval_count": 256,
      "eval_duration_ns": 24107740000,
      "decode_tok_s": 10.619,
      "prompt_eval_count": 505,
      "prompt_eval_duration_ns": 96190000,
      "prefill_tok_s": 5250.026,
      "load_duration_ns": 2730908,
      "total_duration_ns": 24212027532,
      "streamed_chunks": 256,
      "response_chars": 1057,
      "thinking_chars": 0,
      "done_reason": "length",
      "run": 1
    },
    {
      "ttft_ms": 100.08,
      "first_token_was_thinking": false,
      "wall_s": 23.7785,
      "eval_count": 256,
      "eval_duration_ns": 23674789000,
      "decode_tok_s": 10.813,
      "prompt_eval_count": 505,
      "prompt_eval_duration_ns": 93432000,
      "prefill_tok_s": 5405.0,
      "load_duration_ns": 1310274,
      "total_duration_ns": 23774957523,
      "streamed_chunks": 256,
      "response_chars": 1057,
      "thinking_chars": 0,
      "done_reason": "length",
      "run": 2
    },
    {
      "ttft_ms": 99.47,
      "first_token_was_thinking": false,
      "wall_s": 23.1892,
      "eval_count": 256,
      "eval_duration_ns": 23086089000,
      "decode_tok_s": 11.089,
      "prompt_eval_count": 505,
      "prompt_eval_duration_ns": 91484000,
      "prefill_tok_s": 5520.091,
      "load_duration_ns": 1186504,
      "total_duration_ns": 23184892126,
      "streamed_chunks": 256,
      "response_chars": 1057,
      "thinking_chars": 0,
      "done_reason": "length",
      "run": 3
    },
    {
      "ttft_ms": 98.05,
      "first_token_was_thinking": false,
      "wall_s": 22.9416,
      "eval_count": 256,
      "eval_duration_ns": 22840464000,
      "decode_tok_s": 11.208,
      "prompt_eval_count": 505,
      "prompt_eval_duration_ns": 90057000,
      "prefill_tok_s": 5607.56,
      "load_duration_ns": 1621033,
      "total_duration_ns": 22937977809,
      "streamed_chunks": 256,
      "response_chars": 1057,
      "thinking_chars": 0,
      "done_reason": "length",
      "run": 4
    },
    {
      "ttft_ms": 96.18,
      "first_token_was_thinking": false,
      "wall_s": 22.7882,
      "eval_count": 256,
      "eval_duration_ns": 22688727000,
      "decode_tok_s": 11.283,
      "prompt_eval_count": 505,
      "prompt_eval_duration_ns": 89029000,
      "prefill_tok_s": 5672.309,
      "load_duration_ns": 1230454,
      "total_duration_ns": 22784442403,
      "streamed_chunks": 256,
      "response_chars": 1057,
      "thinking_chars": 0,
      "done_reason": "length",
      "run": 5
    }
  ],
  "errors": [],
  "quantization": "Q4_K_M",
  "param_size": "14.7B",
  "family": "phi3",
  "context_length": 16384,
  "think_mode": false,
  "think_note": "think:false accepted",
  "warmup_discarded": {
    "ttft_ms": 109.01,
    "first_token_was_thinking": false,
    "wall_s": 25.0685,
    "eval_count": 256,
    "eval_duration_ns": 24956186000,
    "decode_tok_s": 10.258,
    "prompt_eval_count": 505,
    "prompt_eval_duration_ns": 100522000,
    "prefill_tok_s": 5023.776,
    "load_duration_ns": 1489933,
    "total_duration_ns": 25064209949,
    "streamed_chunks": 256,
    "response_chars": 1057,
    "thinking_chars": 0,
    "done_reason": "length"
  },
  "ps_after_load": [
    {
      "name": "phi4:14b",
      "model": "phi4:14b",
      "size": 10034589201,
      "digest": "ac896e5b8b34a1f4efa7b14d7520725140d5512484457fab45d2a4ea14c69dba",
      "details": {
        "parent_model": "",
        "format": "gguf",
        "family": "phi3",
        "families": [
          "phi3"
        ],
        "parameter_size": "14.7B",
        "quantization_level": "Q4_K_M"
      },
      "expires_at": "2026-08-25T15:22:29.866224577Z",
      "size_vram": 0,
      "context_length": 4096
    }
  ],
  "size_vram": 0,
  "size_total": 10034589201,
  "cpu_only_receipt": true,
  "peak_runner_rss_kb": 9853764,
  "peak_runner_rss_gib": 9.4,
  "peak_runner_threads": 103,
  "rss_samples": 213,
  "loadavg_during_runs": {
    "min": 24.06,
    "median": 28.98,
    "max": 31.0,
    "mean": 28.54,
    "samples": 213,
    "nproc": 64
  },
  "contention_suspected": false,
  "contention_note": "loadavg consistent with this bench owning the machine",
  "foreign_gpu_procs_during": "0, 38764 MiB, 0 %",
  "derived": {
    "decode_tok_s": {
      "min": 10.619,
      "median": 11.089,
      "max": 11.283,
      "mean": 11.002,
      "n": 5
    },
    "ttft_ms": {
      "min": 96.18,
      "median": 99.47,
      "max": 105.23,
      "mean": 99.802,
      "n": 5
    },
    "wall_s": {
      "min": 22.7882,
      "median": 23.1892,
      "max": 24.2159,
      "mean": 23.383,
      "n": 5
    },
    "prefill_tok_s": {
      "min": 5250.026,
      "median": 5520.091,
      "max": 5672.309,
      "mean": 5490.997,
      "n": 5
    },
    "eval_count": {
      "min": 256,
      "median": 256,
      "max": 256,
      "mean": 256.0,
      "n": 5
    }
  },
  "cold_load": [
    {
      "ttft_ms": 19744.93,
      "first_token_was_thinking": false,
      "wall_s": 19.7481,
      "eval_count": 1,
      "eval_duration_ns": 1000,
      "decode_tok_s": 1000000.0,
      "prompt_eval_count": 505,
      "prompt_eval_duration_ns": 8113330000,
      "prefill_tok_s": 62.243,
      "load_duration_ns": 11622819938,
      "total_duration_ns": 19744506361,
      "streamed_chunks": 1,
      "response_chars": 2,
      "thinking_chars": 0,
      "done_reason": "length",
      "wall_to_first_token_s": 19.7482,
      "ps_confirmed_empty_before": true,
      "load_duration_s": 11.6228
    },
    {
      "ttft_ms": 20010.93,
      "first_token_was_thinking": false,
      "wall_s": 20.0141,
      "eval_count": 1,
      "eval_duration_ns": 1000,
      "decode_tok_s": 1000000.0,
      "prompt_eval_count": 505,
      "prompt_eval_duration_ns": 8098697000,
      "prefill_tok_s": 62.356,
      "load_duration_ns": 11903081396,
      "total_duration_ns": 20010324334,
      "streamed_chunks": 1,
      "response_chars": 2,
      "thinking_chars": 0,
      "done_reason": "length",
      "wall_to_first_token_s": 20.0143,
      "ps_confirmed_empty_before": true,
      "load_duration_s": 11.9031
    }
  ],
  "cold_load_note": "PAGE-CACHE-WARM: no sudo to drop caches, so this is warm-disk reload, not a true cold read from NVMe.",
  "nvidia_after": "0, 38764 MiB, 0 %",
  "ps_final": [
    {
      "name": "phi4:14b",
      "model": "phi4:14b",
      "size": 10034589201,
      "digest": "ac896e5b8b34a1f4efa7b14d7520725140d5512484457fab45d2a4ea14c69dba",
      "details": {
        "parent_model": "",
        "format": "gguf",
        "family": "phi3",
        "families": [
          "phi3"
        ],
        "parameter_size": "14.7B",
        "quantization_level": "Q4_K_M"
      },
      "expires_at": "2026-08-25T15:25:09.28581557Z",
      "size_vram": 0,
      "context_length": 4096
    }
  ],
  "finished_utc": "2026-08-25T15:15:10Z"
}