Gemma 4 12B IT benchmark on an M5 Pro · 64 GB

{
  "runId": "run_cc7c5b2a-7d14-47f1-a176-4b5191b0ac62",
  "bundleId": "llamacpp-gemma-4-12b-it-q8_0.gguf-2bb23f",
  "status": "verified",
  "promptTokens": 40960,
  "completionTokens": 10240,
  "contextLength": 5120,
  "harness": {
    "version": "0.1.22",
    "gitSha": "50c24e0"
  },
  "runtime": {
    "name": "llama.cpp",
    "version": "b8920",
    "buildFlags": "metal"
  },
  "model": {
    "displayName": "Gemma 4 12B IT",
    "format": "gguf",
    "quant": "q8_0",
    "architecture": "gemma4",
    "source": "unsloth/gemma-4-12b-it-GGUF:gemma-4-12b-it-Q8_0.gguf",
    "fileSizeBytes": 12669646240,
    "lab": {
      "name": "Google",
      "slug": "google"
    },
    "quantizedBy": {
      "name": "Unsloth",
      "slug": "unsloth"
    }
  },
  "device": {
    "cpu": "Apple M5 Pro",
    "cpuCores": 18,
    "gpu": "Apple M5 Pro",
    "gpuCores": 20,
    "gpuCount": 1,
    "ramGb": 64,
    "osName": "macOS",
    "osVersion": "26.5.1"
  },
  "decodeTpsMean": 21,
  "prefillTpsMean": 620.2,
  "ttftP50Ms": 6715.55,
  "idleTpsMean": 7548.3,
  "peakRssMb": 7643.9,
  "trialsPassed": 10,
  "trialsTotal": 10,
  "runnabilityScore": 0.6365328609793527,
  "bundleSha256": "61a71234f4091a94454bdea14c56983fc43c7280456140596969eae0b321116c",
  "createdAt": "2026-06-06T16:18:43.968Z"
}