Benchmark local LLM models — speed, quality & hardware fitness verdict from any MCP client

{
  "mcpServers": {
    "iogithubmetrillmmetrillm": {
      "command": "npx",
      "args": [
        "-y",
        "metrillm"
      ]
    }
  }
}