{
  "evidenceVersion": "local-ollama-speed-v1",
  "capturedAt": "2026-06-01T21:52:56.274Z",
  "hardwareSlug": "rtx-3080-16gb-mobile",
  "modelTag": "mistral:7b",
  "catalogSlug": "mistral-7b-instruct-v0.3",
  "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=mistral:7b 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": 1873,
      "tempC": 61,
      "smClockMhz": 210,
      "powerDrawW": 19.03
    },
    "ollamaVersionCli": "ollama version is 0.24.0",
    "ollamaApiVersion": {
      "version": "0.24.0"
    }
  },
  "ollamaTagDetails": {
    "name": "mistral:7b",
    "model": "mistral:7b",
    "modified_at": "2026-06-02T00:52:31.00638+03:00",
    "size": 4372824384,
    "digest": "6577803aa9a036369e481d648a2baebb381ebc6e897f2bb9a766a2aa7bfbc1cf",
    "details": {
      "parent_model": "",
      "format": "gguf",
      "family": "llama",
      "families": [
        "llama"
      ],
      "parameter_size": "7.2B",
      "quantization_level": "Q4_K_M"
    }
  },
  "notes": null,
  "logHashSha256": "399886c632ff8c6bfabf8c1cfa73757dc6a5155a674404851fb599544e647929",
  "runsCaptured": 5,
  "decodeStatsAllRuns": {
    "min": 87.67011298109102,
    "max": 91.35543487756374,
    "median": 90.12809067618214,
    "p5": 87.96703409314075,
    "p95": 91.12907295136824,
    "mean": 89.70639646455257,
    "n": 5
  },
  "decodeStatsSteadyRuns": {
    "min": 87.67011298109102,
    "max": 91.35543487756374,
    "median": 89.6414046087609,
    "p5": 87.89280381512832,
    "p95": 91.1713332473565,
    "mean": 89.57708926904414,
    "n": 4
  },
  "runs": [
    {
      "runIndex": 1,
      "startedAt": "2026-06-01T21:52:35.162Z",
      "finishedAt": "2026-06-01T21:52:41.149Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/mistral-7b/run1.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/mistral-7b/run1.log",
      "decodeTokPerSec": 90.22362524658628,
      "prefillTokPerSec": 912.1091612244154,
      "totalMs": 5905.644,
      "loadMs": 2359.7136,
      "promptEvalMs": 16.4454,
      "evalMs": 2837.3943,
      "promptTokens": 15,
      "responseTokens": 256,
      "responseHashSha256": "02e0b2191fe7b698ee839ea5c8577fa05fa4ba3739dd9e4a40a13ea38b5a1796",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 1873,
        "tempC": 61,
        "smClockMhz": 210,
        "powerDrawW": 18.52
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6854,
        "tempC": 67,
        "smClockMhz": 1545,
        "powerDrawW": 114.8
      }
    },
    {
      "runIndex": 2,
      "startedAt": "2026-06-01T21:52:41.223Z",
      "finishedAt": "2026-06-01T21:52:44.911Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/mistral-7b/run2.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/mistral-7b/run2.log",
      "decodeTokPerSec": 91.35543487756374,
      "prefillTokPerSec": 1019.1879110724574,
      "totalMs": 3592.6701,
      "loadMs": 56.0684,
      "promptEvalMs": 14.7176,
      "evalMs": 2802.2416,
      "promptTokens": 15,
      "responseTokens": 256,
      "responseHashSha256": "9571b13f34399799807c25a4ead65fde63a3f9f27e865ce9241d90220891f334",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6854,
        "tempC": 67,
        "smClockMhz": 1545,
        "powerDrawW": 109.11
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6854,
        "tempC": 68,
        "smClockMhz": 1185,
        "powerDrawW": 112.5
      }
    },
    {
      "runIndex": 3,
      "startedAt": "2026-06-01T21:52:44.989Z",
      "finishedAt": "2026-06-01T21:52:48.681Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/mistral-7b/run3.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/mistral-7b/run3.log",
      "decodeTokPerSec": 90.12809067618214,
      "prefillTokPerSec": 994.6619807035576,
      "totalMs": 3620.8929,
      "loadMs": 70.8246,
      "promptEvalMs": 15.0805,
      "evalMs": 2840.4019,
      "promptTokens": 15,
      "responseTokens": 256,
      "responseHashSha256": "9571b13f34399799807c25a4ead65fde63a3f9f27e865ce9241d90220891f334",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6854,
        "tempC": 67,
        "smClockMhz": 1185,
        "powerDrawW": 105.94
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6854,
        "tempC": 69,
        "smClockMhz": 1215,
        "powerDrawW": 110.83
      }
    },
    {
      "runIndex": 4,
      "startedAt": "2026-06-01T21:52:48.755Z",
      "finishedAt": "2026-06-01T21:52:52.476Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/mistral-7b/run4.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/mistral-7b/run4.log",
      "decodeTokPerSec": 89.15471854133965,
      "prefillTokPerSec": 1022.3694434220749,
      "totalMs": 3640.0006,
      "loadMs": 62.455,
      "promptEvalMs": 14.6718,
      "evalMs": 2871.4128,
      "promptTokens": 15,
      "responseTokens": 256,
      "responseHashSha256": "9571b13f34399799807c25a4ead65fde63a3f9f27e865ce9241d90220891f334",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6854,
        "tempC": 69,
        "smClockMhz": 1215,
        "powerDrawW": 110.83
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6854,
        "tempC": 70,
        "smClockMhz": 1155,
        "powerDrawW": 110.97
      }
    },
    {
      "runIndex": 5,
      "startedAt": "2026-06-01T21:52:52.552Z",
      "finishedAt": "2026-06-01T21:52:56.272Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/mistral-7b/run5.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/mistral-7b/run5.log",
      "decodeTokPerSec": 87.67011298109102,
      "prefillTokPerSec": 984.4134536505333,
      "totalMs": 3639.4968,
      "loadMs": 67.6296,
      "promptEvalMs": 15.2375,
      "evalMs": 2920.0373,
      "promptTokens": 15,
      "responseTokens": 256,
      "responseHashSha256": "9571b13f34399799807c25a4ead65fde63a3f9f27e865ce9241d90220891f334",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6854,
        "tempC": 70,
        "smClockMhz": 1155,
        "powerDrawW": 110.97
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 6854,
        "tempC": 71,
        "smClockMhz": 1155,
        "powerDrawW": 110.72
      }
    }
  ],
  "errors": [],
  "uploadEligible": true
}