{
  "protocolVersion": "osai-ctxfit/1",
  "createdAt": "2026-10-20T06:56:29.342Z",
  "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",
    "digest": "359d7dd4bcdab3d86b87d73ac27966f4dbb9f5efdfcc75d34a8764a09474fae7",
    "weightsBytes": 2497293931,
    "family": "qwen3",
    "parameterSize": "4.0B",
    "quantization": "Q4_K_M",
    "trainedContextTokens": 262144
  },
  "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": 1961885696,
        "freeBytes": 10916724736,
        "utilizationPercent": 27
      },
      "gpuAfterLoad": {
        "usedBytes": 5351931904,
        "freeBytes": 7526678528
      },
      "loadDurationNs": 2548529400,
      "ps": {
        "sizeBytes": 3178149969,
        "sizeVramBytes": 3178149969,
        "hostBytes": 0,
        "contextLength": 4096
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 520881999,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 63949000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 512982000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8571000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 503210000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8321000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 498700000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 7902000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 127.18348204526936,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 8192,
      "gpuBefore": {
        "usedBytes": 1944059904,
        "freeBytes": 10934550528,
        "utilizationPercent": 27
      },
      "gpuAfterLoad": {
        "usedBytes": 6030360576,
        "freeBytes": 6848249856
      },
      "loadDurationNs": 1799443300,
      "ps": {
        "sizeBytes": 3873366343,
        "sizeVramBytes": 3873366343,
        "hostBytes": 0,
        "contextLength": 8192
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 518683000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 45460000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 496281000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8330999
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 510304000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8214000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 506761000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8404000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 126.29227584601024,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 16384,
      "gpuBefore": {
        "usedBytes": 1966080000,
        "freeBytes": 10912530432,
        "utilizationPercent": 26
      },
      "gpuAfterLoad": {
        "usedBytes": 7271874560,
        "freeBytes": 5606735872
      },
      "loadDurationNs": 1793157400,
      "ps": {
        "sizeBytes": 5098103111,
        "sizeVramBytes": 5098103111,
        "hostBytes": 0,
        "contextLength": 16384
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 523129000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 42499000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 509879000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8385000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 523537000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8617000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 510339000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8997000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 125.40683741591374,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 32768,
      "gpuBefore": {
        "usedBytes": 2391801856,
        "freeBytes": 10486808576,
        "utilizationPercent": 22
      },
      "gpuAfterLoad": {
        "usedBytes": 10191110144,
        "freeBytes": 2687500288
      },
      "loadDurationNs": 2079780000,
      "ps": {
        "sizeBytes": 7547576647,
        "sizeVramBytes": 7547576647,
        "hostBytes": 0,
        "contextLength": 32768
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 527048000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 42788000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 589358000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 9373000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 557624000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8820000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 559330000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 10116000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 114.42261276884844,
        "fullyInVram": true
      }
    }
  ]
}
