{
  "protocolVersion": "osai-ctxfit/1",
  "createdAt": "2026-10-20T06:30:54.873Z",
  "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": "qwen3:4b-fp16",
    "digest": "c2a4b6294479e77cbc67fe7f72109563831ae07c7ebf5888c410475e9001370b",
    "weightsBytes": 8051298337,
    "family": "qwen3",
    "parameterSize": "4.0B",
    "quantization": "F16",
    "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": 1514143744,
        "freeBytes": 11364466688,
        "utilizationPercent": 22
      },
      "gpuAfterLoad": {
        "usedBytes": 10588520448,
        "freeBytes": 2290089984
      },
      "loadDurationNs": 6115300200,
      "ps": {
        "sizeBytes": 8732152954,
        "sizeVramBytes": 8732152954,
        "hostBytes": 0,
        "contextLength": 4096
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 1339404000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 103063000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 1439467000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 23036000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 1406431000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 22742000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 1295592000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 24419000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 45.505254079297174,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 8192,
      "gpuBefore": {
        "usedBytes": 1653604352,
        "freeBytes": 11225006080,
        "utilizationPercent": 23
      },
      "gpuAfterLoad": {
        "usedBytes": 11167334400,
        "freeBytes": 1711276032
      },
      "loadDurationNs": 5517720000,
      "ps": {
        "sizeBytes": 9340327034,
        "sizeVramBytes": 9340327034,
        "hostBytes": 0,
        "contextLength": 8192
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 1277410000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 101265000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 1267573000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 20769000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 1301652000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 20763000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 1287171000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 20672000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 49.721443382425484,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 16384,
      "gpuBefore": {
        "usedBytes": 1566572544,
        "freeBytes": 11312037888,
        "utilizationPercent": 25
      },
      "gpuAfterLoad": {
        "usedBytes": 12215910400,
        "freeBytes": 662700032
      },
      "loadDurationNs": 6264114300,
      "ps": {
        "sizeBytes": 11419611298,
        "sizeVramBytes": 10344831385,
        "hostBytes": 1074779913,
        "contextLength": 16384
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 2267652000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 103884000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 2203943000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 38930000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 3862033000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 36048000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 2895727000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 37116000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 22.10153097995771,
        "fullyInVram": false
      }
    },
    {
      "numCtx": 32768,
      "gpuBefore": {
        "usedBytes": 1565523968,
        "freeBytes": 11313086464,
        "utilizationPercent": 24
      },
      "gpuAfterLoad": {
        "usedBytes": 12065964032,
        "freeBytes": 812646400
      },
      "loadDurationNs": 8038028200,
      "ps": {
        "sizeBytes": 13953495202,
        "sizeVramBytes": 10442202152,
        "hostBytes": 3511293050,
        "contextLength": 32768
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 9212553000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 178125000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 9408087000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 132692000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 10054359000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 232653000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 9429842000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 128553000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 6.78696419303738,
        "fullyInVram": false
      }
    }
  ]
}
