Gemma 4 31B IT benchmark on an M4 Max · 64 GB

{
  "runId": "run_9400179f-bd64-4a81-8925-934481f6f9ce",
  "bundleId": "llamacpp-gemma-4-31b-it-qat-q4_0.gguf-b56d99",
  "status": "verified",
  "promptTokens": 20480,
  "completionTokens": 5120,
  "contextLength": 5120,
  "harness": {
    "version": "0.1.21",
    "gitSha": "unknown"
  },
  "runtime": {
    "name": "llama.cpp",
    "version": "b9630",
    "buildFlags": "metal"
  },
  "model": {
    "displayName": "Gemma 4 31B IT",
    "format": "gguf",
    "quant": "q4_0",
    "architecture": "gemma4",
    "source": "google/gemma-4-31B-it-qat-q4_0-gguf:gemma-4-31B_q4_0-it.gguf",
    "fileSizeBytes": 17651000768,
    "lab": {
      "name": "Google",
      "slug": "google"
    },
    "quantizedBy": {
      "name": "Google",
      "slug": "google"
    }
  },
  "device": {
    "cpu": "Apple M4 Max",
    "cpuCores": 16,
    "gpu": "Apple M4 Max",
    "gpuCores": 40,
    "gpuCount": 1,
    "ramGb": 64,
    "osName": "macOS",
    "osVersion": "26.5.1"
  },
  "decodeTpsMean": 16.1,
  "prefillTpsMean": 149.9,
  "ttftP50Ms": 27493.07,
  "idleTpsMean": 19318,
  "peakRssMb": 20155.3,
  "trialsPassed": 5,
  "trialsTotal": 5,
  "runnabilityScore": 0.44056979806082597,
  "bundleSha256": "5e8f3c07112c83533232baf4140c48f16e811542e29301afcd468dc7b257fe62",
  "createdAt": "2026-06-21T00:15:32.135Z"
}