# MetriLLM

> Use this tool when you need to evaluate and compare the performance of local Large Language Models (LLMs) on specific hardware configurations. MetriLLM solves problems related to model optimization, hardware selection, and quality assessment by providing comprehensive benchmarks and a global score. It takes LLM models as input and produces a verdict and global score as output, making it ideal for use cases where model performance and hardware compatibility need to be assessed.

Canonical page: https://skillsregistry.net/skills/metrillm  
JSON: https://api.skillsregistry.net/v1/skills/metrillm

## Description

Runs comprehensive benchmarks against local LLM models via Ollama or LM Studio, measuring tokens per second, time to first token, memory usage, and quality across reasoning, coding, math, and multilingual categories. Produces a global score (0-100) combining hardware fit and quality metrics, with a verdict ranging from Excellent to Not Recommended. Results can be shared to a public leaderboard at metrillm.dev for cross-hardware comparison.

## Trust

- **Trust score (0–1):** 0.50
- **Verification tier:** unverified

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** mcp-remote
- **Runtime environment:** api
- **Category:** iot-hardware
- **Updated:** 2026-04-29

## Source

- **Source listing:** [PulseMCP](https://www.pulsemcp.com/servers/metrillm)
- **Repository:** <https://github.com/metrillm/metrillm/tree/HEAD/mcp>

## Use it

Resolve this record through the SkillsRegistry MCP server (no auth, read-only):

```
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp
```

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_skill",
    "arguments": {
      "slug": "metrillm"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/metrillm` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/metrillm/pull`

---
SkillsRegistry indexes agent skills from public registries and GitHub. Skills we have analysed are scanned with Circle-IR and scored on six dimensions; each listing states its scan coverage. More: https://skillsregistry.net/llms.txt