gpt-oss-20b benchmark on an M2 Max · 32 GB

{
  "runId": "run_223ce615-7a8e-4c1d-9374-f389acf12bef",
  "bundleId": "mlx-gpt-oss-20b-mxfp4-q8-cc4d9e",
  "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": "gpt-oss-20b",
    "format": "mlx",
    "quant": "mxfp4",
    "architecture": "gpt_oss",
    "source": "mlx-community/gpt-oss-20b-MXFP4-Q8",
    "fileSizeBytes": 12076207568,
    "lab": {
      "name": "OpenAI",
      "slug": "openai"
    },
    "quantizedBy": {
      "name": "MLX Community",
      "slug": "mlx-community"
    }
  },
  "device": {
    "cpu": "Apple M2 Max",
    "cpuCores": 12,
    "gpu": "Apple M2 Max",
    "gpuCores": 30,
    "gpuCount": 1,
    "ramGb": 32,
    "osName": "macOS",
    "osVersion": "26.4.1"
  },
  "decodeTpsMean": 57.8,
  "prefillTpsMean": 465.8,
  "ttftP50Ms": 8748.01,
  "idleTpsMean": 3892,
  "peakRssMb": 14336,
  "trialsPassed": 10,
  "trialsTotal": 10,
  "runnabilityScore": 0.5935,
  "bundleSha256": "61cc36db80436141056e154a672b50fbde8b53d98172679954958c46fa7179bd",
  "createdAt": "2026-05-11T23:27:35.187Z"
}