Qwen3.5-4B benchmark on an M4 Pro · 48 GB

{
  "runId": "run_7da2d9d8-e7fa-4994-b02a-5a17d6bb0c4d",
  "bundleId": "mlx-qwen3.5-4b-4bit-6e4cdf",
  "status": "verified",
  "promptTokens": 40960,
  "completionTokens": 10240,
  "contextLength": 5120,
  "harness": {
    "version": "0.1.21",
    "gitSha": "unknown"
  },
  "runtime": {
    "name": "mlx_lm",
    "version": "0.31.3",
    "buildFlags": null
  },
  "model": {
    "displayName": "Qwen3.5-4B",
    "format": "mlx",
    "quant": "4bit",
    "architecture": "qwen3_5",
    "source": "mlx-community/Qwen3.5-4B-4bit",
    "fileSizeBytes": 3034300695,
    "lab": {
      "name": "Qwen",
      "slug": "qwen"
    },
    "quantizedBy": {
      "name": "MLX Community",
      "slug": "mlx-community"
    }
  },
  "device": {
    "cpu": "Apple M4 Pro",
    "cpuCores": 14,
    "gpu": "Apple M4 Pro",
    "gpuCores": 20,
    "gpuCount": 1,
    "ramGb": 48,
    "osName": "macOS",
    "osVersion": "26.5.1"
  },
  "decodeTpsMean": 72.3,
  "prefillTpsMean": 619.9,
  "ttftP50Ms": 6687.15,
  "idleTpsMean": 2337,
  "peakRssMb": 4792,
  "trialsPassed": 10,
  "trialsTotal": 10,
  "runnabilityScore": 0.778657075892857,
  "bundleSha256": "6826a1037d561ac702b4b47b830ef2df980e8a6ac169a1b09ea07777ebef26f1",
  "createdAt": "2026-06-30T23:35:02.513Z"
}