{
  "schemaVersion": "0.1",
  "rigrouteLabsVersion": "0.1.0",
  "run": {
    "id": "20260927T165914Z-d4d62449",
    "startedAt": "2026-09-27T16:59:14.790462+00:00",
    "finishedAt": "2026-09-27T17:00:32.226414+00:00"
  },
  "workload": {
    "type": "llm-inference",
    "manifest": "standard-ai-v0.1",
    "modelClass": "8b-q4"
  },
  "hardware": {
    "host": {
      "os": "macOS",
      "osVersion": "26.6.2",
      "kernel": "Darwin 25.6.0",
      "arch": "arm64",
      "machineId": "1f14d1e6bce85456"
    },
    "cpu": {
      "manufacturer": "Apple",
      "model": "Apple M1 Pro",
      "physicalCores": 10,
      "logicalCores": 10
    },
    "memory": {
      "installedBytes": 34359738368,
      "availableAtStartBytes": 17510825984
    },
    "accelerators": [
      {
        "manufacturer": "Apple",
        "model": "Apple M1 Pro",
        "deviceId": "UNKNOWN",
        "memoryArchitecture": "unified-memory",
        "dedicatedVramBytes": null,
        "unifiedMemoryBytes": 34359738368,
        "driver": "UNKNOWN",
        "backend": "Metal"
      }
    ]
  },
  "environment": {
    "drivers": {},
    "runtime": {
      "llamaCppVersion": "0.5.0",
      "llamaBenchBuildNumber": "11146",
      "llamaCppCommit": "7fe450e19",
      "backendsCompiled": "BLAS,CPU,MTL"
    }
  },
  "benchmark": {
    "engine": "llama-bench",
    "engineVersion": "0.5.0 (build 11146, commit 7fe450e19)",
    "model": {
      "path": "<local-path-redacted>/sha256-6a0746a1ec1aef3e7ec53868f220ff6e389f6f8ef87a01d77c96807de94ca2aa",
      "filename": "sha256-6a0746a1ec1aef3e7ec53868f220ff6e389f6f8ef87a01d77c96807de94ca2aa",
      "sizeBytes": 4661211424,
      "sha256": "6a0746a1ec1aef3e7ec53868f220ff6e389f6f8ef87a01d77c96807de94ca2aa",
      "declaredIdentity": {
        "modelType": "llama 8B Q4_0",
        "paramCount": 8030261248,
        "source": "llama-bench JSON output (model_type / model_n_params)"
      },
      "manifestModelClass": "8b-q4"
    },
    "context": 8192,
    "promptTokensPerTest": 512,
    "generationTokensPerTest": 128,
    "repetitions": 5,
    "command": [
      "/opt/homebrew/bin/llama-bench",
      "-m",
      "<local-path-redacted>/sha256-6a0746a1ec1aef3e7ec53868f220ff6e389f6f8ef87a01d77c96807de94ca2aa",
      "-d",
      "8192",
      "-p",
      "512",
      "-n",
      "128",
      "-r",
      "5",
      "-o",
      "json",
      "-v"
    ],
    "deviceUsed": "BLAS,MTL",
    "backendUsed": "BLAS,MTL"
  },
  "measurements": {
    "promptTokensPerSecond": {
      "unit": "tok/s",
      "mean": 190.39043,
      "stddev": 0.274684,
      "samples": [
        189.95,
        190.627,
        190.424,
        190.61,
        190.34
      ]
    },
    "generationTokensPerSecond": {
      "unit": "tok/s",
      "mean": 27.915013,
      "stddev": 0.232568,
      "samples": [
        27.4993,
        28.0319,
        28.0152,
        28.0057,
        28.0229
      ]
    },
    "durationSeconds": 36.37409475,
    "memory": {
      "acceleratorBytes": "UNAVAILABLE",
      "systemBytes": {
        "value": 2568192000.0,
        "unit": "bytes",
        "scope": "system-wall",
        "source": "platform memory snapshot delta (before vs. after benchmark)"
      }
    },
    "power": "UNAVAILABLE",
    "temperature": "UNAVAILABLE"
  },
  "status": "PASS",
  "failure": null,
  "rawEvidence": {
    "command": [
      "/opt/homebrew/bin/llama-bench",
      "-m",
      "<local-path-redacted>/sha256-6a0746a1ec1aef3e7ec53868f220ff6e389f6f8ef87a01d77c96807de94ca2aa",
      "-d",
      "8192",
      "-p",
      "512",
      "-n",
      "128",
      "-r",
      "5",
      "-o",
      "json",
      "-v"
    ],
    "exitCode": 0,
    "stdoutPath": "stdout.log",
    "stderrPath": "stderr.log",
    "llamaBenchJson": [
      {
        "build_commit": "7fe450e19",
        "build_number": 11146,
        "cpu_info": "Accelerate, Apple M1 Pro",
        "gpu_info": "Apple M1 Pro",
        "backends": "BLAS,MTL",
        "model_filename": "<local-path-redacted>/sha256-6a0746a1ec1aef3e7ec53868f220ff6e389f6f8ef87a01d77c96807de94ca2aa",
        "model_type": "llama 8B Q4_0",
        "model_size": 4653375488,
        "model_n_params": 8030261248,
        "n_batch": 2048,
        "n_ubatch": 512,
        "n_threads": 8,
        "cpu_mask": "0x0",
        "cpu_strict": false,
        "poll": 50,
        "type_k": "f16",
        "type_v": "f16",
        "n_gpu_layers": -1,
        "n_cpu_moe": 0,
        "split_mode": "layer",
        "main_gpu": 0,
        "no_kv_offload": false,
        "flash_attn": -1,
        "devices": "auto",
        "tensor_split": "0.00",
        "tensor_buft_overrides": "none",
        "load_mode": "auto",
        "lazy_mode": "auto",
        "embeddings": false,
        "no_op_offload": 0,
        "no_host": false,
        "fit_target": 0,
        "fit_min_ctx": 0,
        "n_prompt": 512,
        "n_gen": 0,
        "n_depth": 8192,
        "test_time": "2026-09-27T16:59:18Z",
        "avg_ns": 2689215283,
        "stddev_ns": 3884331,
        "avg_ts": 190.39043,
        "stddev_ts": 0.274684,
        "samples_ns": [
          2695439708,
          2685869166,
          2688734834,
          2686106417,
          2689926292
        ],
        "samples_ts": [
          189.95,
          190.627,
          190.424,
          190.61,
          190.34
        ]
      },
      {
        "build_commit": "7fe450e19",
        "build_number": 11146,
        "cpu_info": "Accelerate, Apple M1 Pro",
        "gpu_info": "Apple M1 Pro",
        "backends": "BLAS,MTL",
        "model_filename": "<local-path-redacted>/sha256-6a0746a1ec1aef3e7ec53868f220ff6e389f6f8ef87a01d77c96807de94ca2aa",
        "model_type": "llama 8B Q4_0",
        "model_size": 4653375488,
        "model_n_params": 8030261248,
        "n_batch": 2048,
        "n_ubatch": 512,
        "n_threads": 8,
        "cpu_mask": "0x0",
        "cpu_strict": false,
        "poll": 50,
        "type_k": "f16",
        "type_v": "f16",
        "n_gpu_layers": -1,
        "n_cpu_moe": 0,
        "split_mode": "layer",
        "main_gpu": 0,
        "no_kv_offload": false,
        "flash_attn": -1,
        "devices": "auto",
        "tensor_split": "0.00",
        "tensor_buft_overrides": "none",
        "load_mode": "auto",
        "lazy_mode": "auto",
        "embeddings": false,
        "no_op_offload": 0,
        "no_host": false,
        "fit_target": 0,
        "fit_min_ctx": 0,
        "n_prompt": 0,
        "n_gen": 128,
        "n_depth": 8192,
        "test_time": "2026-09-27T17:00:08Z",
        "avg_ns": 4585603666,
        "stddev_ns": 38634094,
        "avg_ts": 27.915013,
        "stddev_ts": 0.232568,
        "samples_ns": [
          4654656833,
          4566227583,
          4568949083,
          4570498667,
          4567686167
        ],
        "samples_ts": [
          27.4993,
          28.0319,
          28.0152,
          28.0057,
          28.0229
        ]
      }
    ]
  }
}