{
  "evidenceVersion": "local-ollama-speed-v1",
  "capturedAt": "2026-06-01T21:38:48.104Z",
  "hardwareSlug": "rtx-3080-16gb-mobile",
  "modelTag": "alibayram/kumru:latest",
  "catalogSlug": "kumru-2b",
  "quant": "Q4_K_M",
  "contextSize": 4096,
  "numPredict": 256,
  "prompt": "Write a detailed explanation of how transformer attention works.",
  "scenario": "single-stream",
  "runtime": "ollama-api",
  "command": "POST http://localhost:11434/api/generate model=alibayram/kumru:latest temperature=0 top_p=1 seed=42 num_ctx=4096 num_predict=256",
  "env": {
    "os": "Microsoft Windows [Version 10.0.26200.8457]",
    "platform": "win32",
    "arch": "x64",
    "cpu": "AMD Ryzen 9 5900HX with Radeon Graphics        ",
    "cpuThreads": 16,
    "ramGb": 31.9,
    "gpu": {
      "vendor": "nvidia",
      "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
      "driver": "571.96",
      "memoryTotalMb": 16384,
      "memoryUsedMb": 4861,
      "tempC": 69,
      "smClockMhz": 825,
      "powerDrawW": 29.48
    },
    "ollamaVersionCli": "ollama version is 0.24.0",
    "ollamaApiVersion": {
      "version": "0.24.0"
    }
  },
  "ollamaTagDetails": {
    "name": "alibayram/kumru:latest",
    "model": "alibayram/kumru:latest",
    "modified_at": "2026-06-02T00:38:34.7908415+03:00",
    "size": 1458559332,
    "digest": "04f3c85761dd4f928a33a739d47924674f2b034d511664213b3576934c7078e6",
    "details": {
      "parent_model": "",
      "format": "gguf",
      "family": "llama",
      "families": [
        "llama"
      ],
      "parameter_size": "2.4B",
      "quantization_level": "Q4_K_M"
    }
  },
  "notes": null,
  "logHashSha256": "7effe7c5bf9fb4d20a8f399280cc1c25565c79d25135b15121636fc85fa37585",
  "runsCaptured": 5,
  "decodeStatsAllRuns": {
    "min": 171.22826111747978,
    "max": 175.79942384485702,
    "median": 174.04572969963922,
    "p5": 171.3602226753246,
    "p95": 175.52160431216052,
    "mean": 173.47436195001086,
    "n": 5
  },
  "decodeStatsSteadyRuns": {
    "min": 171.22826111747978,
    "max": 175.79942384485702,
    "median": 174.22802794050685,
    "p5": 171.6508814048037,
    "p95": 175.59105919533462,
    "mean": 173.87093521083762,
    "n": 4
  },
  "runs": [
    {
      "runIndex": 1,
      "startedAt": "2026-06-01T21:38:36.927Z",
      "finishedAt": "2026-06-01T21:38:40.390Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/alibayram-kumru-latest/run1.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/alibayram-kumru-latest/run1.log",
      "decodeTokPerSec": 171.88806890670392,
      "prefillTokPerSec": 2220.094631533669,
      "totalMs": 3370.7463,
      "loadMs": 1691.1365,
      "promptEvalMs": 14.4138,
      "evalMs": 1489.3413,
      "promptTokens": 32,
      "responseTokens": 256,
      "responseHashSha256": "7d8029004316ddd9ae42f729fc79f20d569b8d2624db50467c67e40728daba69",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 4861,
        "tempC": 69,
        "smClockMhz": 825,
        "powerDrawW": 29.49
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6639,
        "tempC": 74,
        "smClockMhz": 1425,
        "powerDrawW": 114.25
      }
    },
    {
      "runIndex": 2,
      "startedAt": "2026-06-01T21:38:40.462Z",
      "finishedAt": "2026-06-01T21:38:42.292Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/alibayram-kumru-latest/run2.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/alibayram-kumru-latest/run2.log",
      "decodeTokPerSec": 175.79942384485702,
      "prefillTokPerSec": 3867.8640929254343,
      "totalMs": 1728.3959,
      "loadMs": 117.1697,
      "promptEvalMs": 8.2733,
      "evalMs": 1456.205,
      "promptTokens": 32,
      "responseTokens": 256,
      "responseHashSha256": "db877b117259948f0ca7c3471d8fdda31b98e99036131d684b7b7328b310202d",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6639,
        "tempC": 74,
        "smClockMhz": 1425,
        "powerDrawW": 108.36
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6639,
        "tempC": 75,
        "smClockMhz": 1485,
        "powerDrawW": 113.88
      }
    },
    {
      "runIndex": 3,
      "startedAt": "2026-06-01T21:38:42.380Z",
      "finishedAt": "2026-06-01T21:38:44.214Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/alibayram-kumru-latest/run3.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/alibayram-kumru-latest/run3.log",
      "decodeTokPerSec": 174.04572969963922,
      "prefillTokPerSec": 4032.0040320040316,
      "totalMs": 1741.3575,
      "loadMs": 117.4736,
      "promptEvalMs": 7.9365,
      "evalMs": 1470.8778,
      "promptTokens": 32,
      "responseTokens": 256,
      "responseHashSha256": "db877b117259948f0ca7c3471d8fdda31b98e99036131d684b7b7328b310202d",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6639,
        "tempC": 73,
        "smClockMhz": 1485,
        "powerDrawW": 107.64
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6639,
        "tempC": 75,
        "smClockMhz": 1470,
        "powerDrawW": 114.29
      }
    },
    {
      "runIndex": 4,
      "startedAt": "2026-06-01T21:38:44.290Z",
      "finishedAt": "2026-06-01T21:38:46.145Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/alibayram-kumru-latest/run4.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/alibayram-kumru-latest/run4.log",
      "decodeTokPerSec": 171.22826111747978,
      "prefillTokPerSec": 4844.154467975598,
      "totalMs": 1771.2944,
      "loadMs": 126.7029,
      "promptEvalMs": 6.6059,
      "evalMs": 1495.0803,
      "promptTokens": 32,
      "responseTokens": 256,
      "responseHashSha256": "db877b117259948f0ca7c3471d8fdda31b98e99036131d684b7b7328b310202d",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6639,
        "tempC": 75,
        "smClockMhz": 1470,
        "powerDrawW": 109.53
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6630,
        "tempC": 76,
        "smClockMhz": 1470,
        "powerDrawW": 112.16
      }
    },
    {
      "runIndex": 5,
      "startedAt": "2026-06-01T21:38:46.224Z",
      "finishedAt": "2026-06-01T21:38:48.102Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/alibayram-kumru-latest/run5.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/alibayram-kumru-latest/run5.log",
      "decodeTokPerSec": 174.41032618137447,
      "prefillTokPerSec": 4108.885464817668,
      "totalMs": 1790.7268,
      "loadMs": 127.3775,
      "promptEvalMs": 7.788,
      "evalMs": 1467.803,
      "promptTokens": 32,
      "responseTokens": 256,
      "responseHashSha256": "db877b117259948f0ca7c3471d8fdda31b98e99036131d684b7b7328b310202d",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6630,
        "tempC": 76,
        "smClockMhz": 1470,
        "powerDrawW": 106.16
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6630,
        "tempC": 76,
        "smClockMhz": 1470,
        "powerDrawW": 114.46
      }
    }
  ],
  "errors": [],
  "uploadEligible": true
}