Qwen3.6-35B-A3B benchmark on an M4 · 32 GB

{
  "runId": "run_73fa6bf5-a5b7-4d97-9f51-13c8a957e8aa",
  "bundleId": "mlx-qwen3.6-35b-a3b-q3-mlx-53925a",
  "status": "verified",
  "promptTokens": 40960,
  "completionTokens": 10240,
  "contextLength": 5120,
  "harness": {
    "version": "0.1.21",
    "gitSha": "be5e71cb"
  },
  "runtime": {
    "name": "mlx_lm",
    "version": "0.31.2",
    "buildFlags": null
  },
  "model": {
    "displayName": "Qwen3.6-35B-A3B",
    "format": "mlx",
    "quant": "3bit",
    "architecture": "qwen3_5_moe",
    "source": "jedisct1/Qwen3.6-35B-A3B-q3-mlx",
    "fileSizeBytes": 15179224487,
    "lab": {
      "name": "Qwen",
      "slug": "qwen"
    },
    "quantizedBy": {
      "name": "Frank Denis",
      "slug": "jedisct1"
    }
  },
  "device": {
    "cpu": "Apple M4",
    "cpuCores": 10,
    "gpu": "Apple M4",
    "gpuCores": 10,
    "gpuCount": 1,
    "ramGb": 32,
    "osName": "macOS",
    "osVersion": "26.2"
  },
  "decodeTpsMean": 40.2,
  "prefillTpsMean": 305.2,
  "ttftP50Ms": 14242.6,
  "idleTpsMean": 3774,
  "peakRssMb": 17408,
  "trialsPassed": 10,
  "trialsTotal": 10,
  "runnabilityScore": 0.5001547619047619,
  "bundleSha256": "8299ecbe252c25245fb78a7ebdfd3abc108df36cae360fa8617f832391c8c54f",
  "createdAt": "2026-04-22T13:20:07.637Z"
}