{
  "schema_version": 1,
  "status": "complete",
  "sample_count": 12,
  "source_sha256": "da2c8a96da4992146eb4146191c2a1f0ca9934cefe182066341c19f6a1233a4a",
  "rows": [
    {
      "case": "document-1k",
      "prompt_tokens": 840,
      "samples": 3,
      "request_input_tokens": 840,
      "reused_prefix_tokens": 0,
      "ttft_ms": {
        "median": 8167.319291038439,
        "min": 3794.5201250258833,
        "max": 11776.336583076045
      },
      "first_visible_text_ms": {
        "median": 8167.319291038439,
        "min": 3794.5201250258833,
        "max": 11776.336583076045
      },
      "wall_ms": {
        "median": 26801.746750017628,
        "min": 13193.985000019893,
        "max": 34020.97675006371
      },
      "stream_decode_tokens_per_second": {
        "median": 13.702565806122495,
        "min": 11.47938251974066,
        "max": 27.17281209023034
      },
      "backend_generation_tokens_per_second": {
        "median": 13.756274254031888,
        "min": 11.524376252394214,
        "max": 27.27923805751722
      },
      "peak_mlx_allocation_bytes": {
        "median": 16826468714,
        "min": 16826468714,
        "max": 16826468714
      },
      "prefix_prepare_ms_excluded_from_request_timing": {
        "median": 0.0,
        "min": 0.0,
        "max": 0.0
      },
      "backend_generation_tokens": {
        "median": 256,
        "min": 256,
        "max": 256
      }
    },
    {
      "case": "document-4k",
      "prompt_tokens": 3156,
      "samples": 3,
      "request_input_tokens": 3156,
      "reused_prefix_tokens": 0,
      "ttft_ms": {
        "median": 27664.122125017457,
        "min": 16594.743250054307,
        "max": 35354.14279100951
      },
      "first_visible_text_ms": {
        "median": 27664.122125017457,
        "min": 16594.743250054307,
        "max": 35354.14279100951
      },
      "wall_ms": {
        "median": 46414.80004100595,
        "min": 26450.194125063717,
        "max": 58372.908457997255
      },
      "stream_decode_tokens_per_second": {
        "median": 13.616583152942251,
        "min": 11.090324464268846,
        "max": 25.921675566759685
      },
      "backend_generation_tokens_per_second": {
        "median": 13.669944107421376,
        "min": 11.133798097291741,
        "max": 26.023246087231495
      },
      "peak_mlx_allocation_bytes": {
        "median": 18206125460,
        "min": 18206125460,
        "max": 18206125460
      },
      "prefix_prepare_ms_excluded_from_request_timing": {
        "median": 0.0,
        "min": 0.0,
        "max": 0.0
      },
      "backend_generation_tokens": {
        "median": 256,
        "min": 256,
        "max": 256
      }
    },
    {
      "case": "document-16k",
      "prompt_tokens": 12263,
      "samples": 3,
      "request_input_tokens": 12263,
      "reused_prefix_tokens": 0,
      "ttft_ms": {
        "median": 113210.93154198024,
        "min": 74392.49441609718,
        "max": 133192.02741701156
      },
      "first_visible_text_ms": {
        "median": 113210.93154198024,
        "min": 74392.49441609718,
        "max": 133192.02741701156
      },
      "wall_ms": {
        "median": 132353.01725007594,
        "min": 88409.49083305895,
        "max": 158487.5147920102
      },
      "stream_decode_tokens_per_second": {
        "median": 13.343624465345108,
        "min": 10.09524497435868,
        "max": 18.21734449243082
      },
      "backend_generation_tokens_per_second": {
        "median": 13.39578002318682,
        "min": 10.13477562438215,
        "max": 18.288754409607378
      },
      "peak_mlx_allocation_bytes": {
        "median": 19840039990,
        "min": 19840039990,
        "max": 19840039990
      },
      "prefix_prepare_ms_excluded_from_request_timing": {
        "median": 0.0,
        "min": 0.0,
        "max": 0.0
      },
      "backend_generation_tokens": {
        "median": 256,
        "min": 256,
        "max": 256
      }
    },
    {
      "case": "document-16k-cached-prefix",
      "prompt_tokens": 12263,
      "samples": 3,
      "request_input_tokens": 64,
      "reused_prefix_tokens": 12199,
      "ttft_ms": {
        "median": 868.3726669987664,
        "min": 710.6026670662686,
        "max": 1340.1490420801565
      },
      "first_visible_text_ms": {
        "median": 868.3726669987664,
        "min": 710.6026670662686,
        "max": 1340.1490420801565
      },
      "wall_ms": {
        "median": 21546.954457997344,
        "min": 18513.289291993715,
        "max": 24455.68362507038
      },
      "stream_decode_tokens_per_second": {
        "median": 12.250760381892167,
        "min": 11.045875828986997,
        "max": 14.478664869233716
      },
      "backend_generation_tokens_per_second": {
        "median": 12.298779043717674,
        "min": 11.089136007539738,
        "max": 14.535384283435764
      },
      "peak_mlx_allocation_bytes": {
        "median": 19717053138,
        "min": 19717053138,
        "max": 19717053138
      },
      "prefix_prepare_ms_excluded_from_request_timing": {
        "median": 120802.1089170361,
        "min": 118645.21641598549,
        "max": 134772.5804169895
      },
      "backend_generation_tokens": {
        "median": 256,
        "min": 256,
        "max": 256
      },
      "output_token_equality_by_repetition": [
        false,
        false,
        false
      ]
    }
  ],
  "model_load_seconds": 2.6883932079654187,
  "loaded_model_mlx_bytes": 15133588488,
  "shared_workstation": true,
  "minimum_observed_memory_free_percent": 23,
  "baseline_system_swap_used_bytes": 13369606144,
  "maximum_observed_additional_system_swap_bytes": 0
}
