{
  "protocolVersion": "osai-ctxfit/1",
  "createdAt": "2026-10-20T06:57:19.158Z",
  "tool": "scripts/lab/context-fit-sweep.mjs",
  "runtime": {
    "name": "ollama",
    "version": "0.35.1",
    "endpoint": "loopback"
  },
  "system": {
    "cpu": {
      "model": "Intel(R) Core(TM) i7-14700K",
      "logicalCores": 28
    },
    "memory": {
      "totalBytes": 34018738176
    },
    "os": {
      "platform": "win32",
      "release": "10.0.26200"
    },
    "gpu": {
      "model": "NVIDIA GeForce RTX 4070 Ti",
      "totalVramBytes": 12878610432,
      "driverVersion": "617.42",
      "provider": "nvidia-smi"
    }
  },
  "model": {
    "identifier": "qwen3:8b",
    "digest": "500a1f067a9f782620b40bee6f7b0c89e17ae61f686b92c24933e4ca4b2b8b41",
    "weightsBytes": 5225388164,
    "family": "qwen3",
    "parameterSize": "8.2B",
    "quantization": "Q4_K_M",
    "trainedContextTokens": 40960
  },
  "settings": {
    "contexts": [
      4096,
      8192,
      16384,
      32768
    ],
    "skippedContexts": [],
    "passes": 4,
    "warmups": 1,
    "numPredict": 64,
    "temperature": 0,
    "seed": 42,
    "prompt": "fixed short prompt, see PROMPT in scripts/lab/context-fit-sweep.mjs",
    "settleSeconds": 8
  },
  "steps": [
    {
      "numCtx": 4096,
      "gpuBefore": {
        "usedBytes": 2022703104,
        "freeBytes": 10855907328,
        "utilizationPercent": 22
      },
      "gpuAfterLoad": {
        "usedBytes": 7830765568,
        "freeBytes": 5047844864
      },
      "loadDurationNs": 3310890000,
      "ps": {
        "sizeBytes": 5578204118,
        "sizeVramBytes": 5578204118,
        "hostBytes": 0,
        "contextLength": 4096
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 906218000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 151693000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 857627000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 14049000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 869715000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 14492000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 872260000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 14325000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 73.58732458334052,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 8192,
      "gpuBefore": {
        "usedBytes": 2020605952,
        "freeBytes": 10858004480,
        "utilizationPercent": 23
      },
      "gpuAfterLoad": {
        "usedBytes": 8514437120,
        "freeBytes": 4364173312
      },
      "loadDurationNs": 2296951700,
      "ps": {
        "sizeBytes": 6295440588,
        "sizeVramBytes": 6295440588,
        "hostBytes": 0,
        "contextLength": 8192
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 882906000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 90017000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 869814000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13391000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 874527000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 14358000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 877134000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 14651000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 73.18241746681349,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 16384,
      "gpuBefore": {
        "usedBytes": 2441084928,
        "freeBytes": 10437525504,
        "utilizationPercent": 25
      },
      "gpuAfterLoad": {
        "usedBytes": 9995026432,
        "freeBytes": 2883584000
      },
      "loadDurationNs": 2303014600,
      "ps": {
        "sizeBytes": 7520177356,
        "sizeVramBytes": 7520177356,
        "hostBytes": 0,
        "contextLength": 16384
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 885773000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 91132000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 872867000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13531000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 917785000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 14285000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 881295000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 14173000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 72.62040519916712,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 32768,
      "gpuBefore": {
        "usedBytes": 2360344576,
        "freeBytes": 10518265856,
        "utilizationPercent": 12
      },
      "gpuAfterLoad": {
        "usedBytes": 11860443136,
        "freeBytes": 1018167296
      },
      "loadDurationNs": 2785476500,
      "ps": {
        "sizeBytes": 9835422678,
        "sizeVramBytes": 9835422678,
        "hostBytes": 0,
        "contextLength": 32768
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 3255105000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 103686000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 3045260000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 51704000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 4016347000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 31062000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 3112565000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 36147000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 20.561819592522564,
        "fullyInVram": true
      }
    }
  ]
}
