Qwen3.8-27B benchmark on an M1 Max · 64 GB

{
  "runId": "run_9cd9bc25-0608-4a03-8d78-91cba506acba",
  "bundleId": "llamacpp-qwen3.8-27b-ud-q4_k_s.gguf-cc2187",
  "status": "verified",
  "promptTokens": 40960,
  "completionTokens": 10240,
  "contextLength": 5120,
  "harness": {
    "version": "0.1.21",
    "gitSha": "159b74142"
  },
  "runtime": {
    "name": "llama.cpp",
    "version": "0.3.0-dev",
    "buildFlags": null
  },
  "model": {
    "displayName": "Qwen3.8-27B",
    "format": "gguf",
    "quant": "q4_k_s",
    "architecture": "qwen35",
    "source": "unsloth/Qwen3.8-27B-GGUF:Qwen3.8-27B-UD-Q4_K_S.gguf",
    "fileSizeBytes": 15358213024,
    "lab": {
      "name": "Qwen",
      "slug": "qwen"
    },
    "quantizedBy": {
      "name": "Unsloth",
      "slug": "unsloth"
    }
  },
  "device": {
    "cpu": "Apple M1 Max",
    "cpuCores": 10,
    "gpu": "Apple M1 Max",
    "gpuCores": 32,
    "gpuCount": 1,
    "ramGb": 64,
    "osName": "macOS",
    "osVersion": "27.0"
  },
  "decodeTpsMean": 8.1,
  "prefillTpsMean": 93.7,
  "ttftP50Ms": 43678.88,
  "idleTpsMean": 15819.2,
  "peakRssMb": 15872.4,
  "trialsPassed": 10,
  "trialsTotal": 10,
  "runnabilityScore": 0.3654277413504464,
  "bundleSha256": "fa72237d426ff7bbaf4595af2ff393a01c9671a2d47caed310c47b24c4c78ced",
  "createdAt": "2026-09-02T23:03:38.262Z"
}