LFM2-1.2B-Extract benchmark on an M1 Max · 64 GB

{
  "runId": "run_8ca209a8-9833-463d-9c15-307cbc932a19",
  "bundleId": "llamacpp-lfm2-1.2b-extract-f16.gguf-25a069",
  "status": "verified",
  "promptTokens": 40960,
  "completionTokens": 10240,
  "contextLength": 5120,
  "harness": {
    "version": "0.1.21",
    "gitSha": "753787f"
  },
  "runtime": {
    "name": "llama.cpp",
    "version": "b8640",
    "buildFlags": "metal"
  },
  "model": {
    "displayName": "LFM2-1.2B-Extract",
    "format": "gguf",
    "quant": "f16",
    "architecture": "lfm2",
    "source": "LiquidAI/LFM2-1.2B-Extract-GGUF:LFM2-1.2B-Extract-F16.gguf",
    "fileSizeBytes": 2343325408,
    "lab": {
      "name": "Liquid AI",
      "slug": "liquid"
    },
    "quantizedBy": {
      "name": "Liquid AI",
      "slug": "liquid"
    }
  },
  "device": {
    "cpu": "Apple M1 Max",
    "cpuCores": 10,
    "gpu": "Apple M1 Max",
    "gpuCores": 32,
    "gpuCount": 1,
    "ramGb": 64,
    "osName": "macOS",
    "osVersion": "15.6.1"
  },
  "decodeTpsMean": 89.2,
  "prefillTpsMean": 2098.8,
  "ttftP50Ms": 2329.19,
  "idleTpsMean": 1782.2,
  "peakRssMb": 2733.1,
  "trialsPassed": 10,
  "trialsTotal": 10,
  "runnabilityScore": 0.918396944405692,
  "bundleSha256": "3f5954f085da57af757371af920061f55baf947301cc0631f3717b3ebac70c87",
  "createdAt": "2026-04-14T03:54:20.775Z"
}