{
  "model_id": "m6",
  "model_tag": "nemotron-3.5-lightning:30b-a3b",
  "box_class": "cpu-server",
  "started_utc": "2026-08-25T15:23:29Z",
  "nproc": 64,
  "num_predict": 256,
  "prompt_sha256": "90eedd0c53f9554ae3837674504fcb7090013d9a432a653352183c0f25a7ce5c",
  "nvidia_before": "0, 38764 MiB, 0 %",
  "ollama_version": "0.32.15",
  "runs": [
    {
      "ttft_ms": 120.95,
      "first_token_was_thinking": false,
      "wall_s": 8.3225,
      "eval_count": 256,
      "eval_duration_ns": 8199133000,
      "decode_tok_s": 31.223,
      "prompt_eval_count": 532,
      "prompt_eval_duration_ns": 115821000,
      "prefill_tok_s": 4593.295,
      "load_duration_ns": 1093545,
      "total_duration_ns": 8319833067,
      "streamed_chunks": 245,
      "response_chars": 1029,
      "thinking_chars": 0,
      "done_reason": "length",
      "run": 1
    },
    {
      "ttft_ms": 114.24,
      "first_token_was_thinking": false,
      "wall_s": 8.3131,
      "eval_count": 256,
      "eval_duration_ns": 8196193999,
      "decode_tok_s": 31.234,
      "prompt_eval_count": 532,
      "prompt_eval_duration_ns": 109087000,
      "prefill_tok_s": 4876.841,
      "load_duration_ns": 1302324,
      "total_duration_ns": 8310393031,
      "streamed_chunks": 245,
      "response_chars": 1029,
      "thinking_chars": 0,
      "done_reason": "length",
      "run": 2
    },
    {
      "ttft_ms": 114.21,
      "first_token_was_thinking": false,
      "wall_s": 8.2327,
      "eval_count": 256,
      "eval_duration_ns": 8115814000,
      "decode_tok_s": 31.543,
      "prompt_eval_count": 532,
      "prompt_eval_duration_ns": 109090000,
      "prefill_tok_s": 4876.707,
      "load_duration_ns": 1253695,
      "total_duration_ns": 8229625381,
      "streamed_chunks": 245,
      "response_chars": 1029,
      "thinking_chars": 0,
      "done_reason": "length",
      "run": 3
    },
    {
      "ttft_ms": 115.0,
      "first_token_was_thinking": false,
      "wall_s": 8.2522,
      "eval_count": 256,
      "eval_duration_ns": 8134976000,
      "decode_tok_s": 31.469,
      "prompt_eval_count": 532,
      "prompt_eval_duration_ns": 109728000,
      "prefill_tok_s": 4848.352,
      "load_duration_ns": 1267734,
      "total_duration_ns": 8249443593,
      "streamed_chunks": 245,
      "response_chars": 1029,
      "thinking_chars": 0,
      "done_reason": "length",
      "run": 4
    },
    {
      "ttft_ms": 113.29,
      "first_token_was_thinking": false,
      "wall_s": 8.1451,
      "eval_count": 256,
      "eval_duration_ns": 8028973000,
      "decode_tok_s": 31.885,
      "prompt_eval_count": 532,
      "prompt_eval_duration_ns": 108356000,
      "prefill_tok_s": 4909.742,
      "load_duration_ns": 1203765,
      "total_duration_ns": 8142352191,
      "streamed_chunks": 245,
      "response_chars": 1029,
      "thinking_chars": 0,
      "done_reason": "length",
      "run": 5
    }
  ],
  "errors": [],
  "quantization": "Q4_K_M",
  "param_size": "32.9B",
  "family": "nemotron_h_moe",
  "context_length": 1048576,
  "think_mode": false,
  "think_note": "think:false accepted",
  "warmup_discarded": {
    "ttft_ms": 113.03,
    "first_token_was_thinking": false,
    "wall_s": 8.2688,
    "eval_count": 256,
    "eval_duration_ns": 8153211000,
    "decode_tok_s": 31.399,
    "prompt_eval_count": 532,
    "prompt_eval_duration_ns": 108189000,
    "prefill_tok_s": 4917.321,
    "load_duration_ns": 1144665,
    "total_duration_ns": 8265871419,
    "streamed_chunks": 245,
    "response_chars": 1029,
    "thinking_chars": 0,
    "done_reason": "length"
  },
  "ps_after_load": [
    {
      "name": "nemotron-3.5-lightning:30b-a3b",
      "model": "nemotron-3.5-lightning:30b-a3b",
      "size": 25609340188,
      "digest": "e7a64ff15fb174c42b4f463e5c888c4f2c7b9cabf9e8d65a1c0874405426c1b2",
      "details": {
        "parent_model": "",
        "format": "gguf",
        "family": "nemotron_h_moe",
        "families": [
          "nemotron_h_moe"
        ],
        "parameter_size": "32.9B",
        "quantization_level": "Q4_K_M"
      },
      "expires_at": "2026-08-25T15:34:01.263085053Z",
      "size_vram": 0,
      "context_length": 4096
    }
  ],
  "size_vram": 0,
  "size_total": 25609340188,
  "cpu_only_receipt": true,
  "peak_runner_rss_kb": 25352424,
  "peak_runner_rss_gib": 24.18,
  "peak_runner_threads": 130,
  "rss_samples": 75,
  "loadavg_during_runs": {
    "min": 19.46,
    "median": 23.02,
    "max": 25.57,
    "mean": 22.83,
    "samples": 75,
    "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": 31.223,
      "median": 31.469,
      "max": 31.885,
      "mean": 31.471,
      "n": 5
    },
    "ttft_ms": {
      "min": 113.29,
      "median": 114.24,
      "max": 120.95,
      "mean": 115.538,
      "n": 5
    },
    "wall_s": {
      "min": 8.1451,
      "median": 8.2522,
      "max": 8.3225,
      "mean": 8.253,
      "n": 5
    },
    "prefill_tok_s": {
      "min": 4593.295,
      "median": 4876.707,
      "max": 4909.742,
      "mean": 4820.987,
      "n": 5
    },
    "eval_count": {
      "min": 256,
      "median": 256,
      "max": 256,
      "mean": 256.0,
      "n": 5
    }
  },
  "cold_load": [
    {
      "ttft_ms": 20502.39,
      "first_token_was_thinking": false,
      "wall_s": 20.5041,
      "eval_count": 1,
      "eval_duration_ns": 1000,
      "decode_tok_s": 1000000.0,
      "prompt_eval_count": 532,
      "prompt_eval_duration_ns": 4025199000,
      "prefill_tok_s": 132.167,
      "load_duration_ns": 16470764633,
      "total_duration_ns": 20501832123,
      "streamed_chunks": 1,
      "response_chars": 4,
      "thinking_chars": 0,
      "done_reason": "length",
      "wall_to_first_token_s": 20.5042,
      "ps_confirmed_empty_before": true,
      "load_duration_s": 16.4708
    },
    {
      "ttft_ms": 20502.19,
      "first_token_was_thinking": false,
      "wall_s": 20.5038,
      "eval_count": 1,
      "eval_duration_ns": 1000,
      "decode_tok_s": 1000000.0,
      "prompt_eval_count": 532,
      "prompt_eval_duration_ns": 4024625000,
      "prefill_tok_s": 132.186,
      "load_duration_ns": 16472207663,
      "total_duration_ns": 20501485370,
      "streamed_chunks": 1,
      "response_chars": 4,
      "thinking_chars": 0,
      "done_reason": "length",
      "wall_to_first_token_s": 20.504,
      "ps_confirmed_empty_before": true,
      "load_duration_s": 16.4722
    }
  ],
  "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": "nemotron-3.5-lightning:30b-a3b",
      "model": "nemotron-3.5-lightning:30b-a3b",
      "size": 25609340188,
      "digest": "e7a64ff15fb174c42b4f463e5c888c4f2c7b9cabf9e8d65a1c0874405426c1b2",
      "details": {
        "parent_model": "",
        "format": "gguf",
        "family": "nemotron_h_moe",
        "families": [
          "nemotron_h_moe"
        ],
        "parameter_size": "32.9B",
        "quantization_level": "Q4_K_M"
      },
      "expires_at": "2026-08-25T15:35:29.311360095Z",
      "size_vram": 0,
      "context_length": 4096
    }
  ],
  "finished_utc": "2026-08-25T15:25:32Z"
}