LFM2.5-VL-450M benchmark on an M1 Max · 64 GB

{
  "runId": "run_c38ff872-d4ea-405a-bc7b-976bc269746f",
  "bundleId": "llamacpp-lfm2.5-vl-450m-q8_0.gguf-3e5413",
  "status": "verified",
  "promptTokens": 40960,
  "completionTokens": 10240,
  "contextLength": 5120,
  "harness": {
    "version": "0.1.20",
    "gitSha": "753787f"
  },
  "runtime": {
    "name": "llama.cpp",
    "version": "b8640",
    "buildFlags": "metal"
  },
  "model": {
    "displayName": "LFM2.5-VL-450M",
    "format": "gguf",
    "quant": "q8_0",
    "architecture": "lfm2",
    "source": "LiquidAI/LFM2.5-VL-450M-GGUF:LFM2.5-VL-450M-Q8_0.gguf",
    "fileSizeBytes": 379219104,
    "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": 256.1,
  "prefillTpsMean": 10490.4,
  "ttftP50Ms": 390.52,
  "idleTpsMean": 381.6,
  "peakRssMb": 475.9,
  "trialsPassed": 10,
  "trialsTotal": 10,
  "runnabilityScore": 0.9968878609793526,
  "bundleSha256": "2f6bfa73f30cd59e6ac80de5443e4464853327d2bacb2b1ab0043fa4b5f66bd5",
  "createdAt": "2026-04-09T04:48:27.236Z"
}