# Model Tested

> Model Tested is a free MCP server for ai models and llm tooling. Independent open-weight LLM test scores: which local model fits your GPU or Mac. It runs as a hosted remote server, so there is nothing to install locally.

- **Kind:** MCP server
- **Pricing:** Free
- **Platforms:** API, Web
- **Website:** https://modeltested.com/mcp/
- **Repository:** https://github.com/jelly-ham/modeltested/tree/HEAD/site/worker
- **MCP Registry name:** com.modeltested/llm-benchmarks
- **Categories:** AI models and LLM tooling
- **Listing updated:** 2026-11-16

## About

Model Tested runs open-weight models on practical work and publishes the scores; the MCP server exposes those numbers to an assistant, covering 70 models as of the November 14, 2026 update. It is free, read-only and needs no account or API key, and the site says it does not log what you ask.

The benchmark uses three tests: coding graded by running the model's code against hidden unit tests in a no-network Docker sandbox, document reading graded field by field against a calculated key, and decisions scored on accuracy and Brier calibration. Most questions are hidden, and the server never returns them or the models' answers. Models run through OpenRouter, with some Q4 downloads also tried on a Mac mini M4 with 16 GB.

Memory estimates assume an 8k-token context. The project deliberately publishes no speed figures, because they depend on hardware.

## Features

- **Fit by hardware:** Finds models that fit a graphics card or Mac of a given memory size.
- **Task rankings:** Best models overall or for coding, documents, decisions and hard tasks.
- **Side-by-side comparison:** Two to six models compared on scores and memory.
- **Decision model rankings:** Accuracy, honesty of confidence and cost per 1,000 calls.
- **No account needed:** Free, read-only, no API key.

## Good for

- Pick a local model for a 16 GB GPU
- Find the best coding model for a 32 GB Mac
- Compare two similar models before downloading
- Check memory needed at a given quantization

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http llm-benchmarks https://modeltested.com/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://modeltested.com/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "llm-benchmarks": {
      "url": "https://modeltested.com/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "llm-benchmarks": {
      "type": "http",
      "url": "https://modeltested.com/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "llm-benchmarks": {
      "serverUrl": "https://modeltested.com/mcp"
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "llm-benchmarks": {
      "type": "http",
      "url": "https://modeltested.com/mcp"
    }
  }
}
```

## Tools

- `find_models`: Best models for a task, optionally limited to those fitting a given GPU or Mac memory.
- `get_model`: All data on one model: scores, rank, memory per quantization and smallest hardware that fits.
- `compare_models`: Two to six models side by side.
- `decision_models`: Decision model rankings with accuracy, hard tasks, calibration and cost per 1,000 calls.
- `search_articles`: Search articles and ranking pages by keyword.

## FAQ

### Is it free?

Yes, free and read-only with no account or key.

### Which clients work?

Claude, ChatGPT, Cursor, VS Code and any app that supports remote MCP over HTTP.

### Does it report speed?

No. The project does not measure speed because it depends on hardware.

### Are the test questions exposed?

No. The server returns only scores, never hidden questions or answers.

### Is Model Tested free?

Yes. Model Tested is free to use.

### How do I add Model Tested to Claude Code?

Run this command in your terminal: claude mcp add --transport http llm-benchmarks https://modeltested.com/mcp

### What are the best alternatives to Model Tested?

The closest alternatives to Model Tested are Claude Code Ultimate Guide, signafAIde, Toolkien, Lugh and HypePaper.

## Alternatives

- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [signafAIde](https://ai-powered.com/mcp/signafaide): signafAIde is a freemium MCP server for security.
- [Toolkien](https://ai-powered.com/mcp/toolkien): Toolkien is an open-source MCP server for ai models and llm tooling.
- [Lugh](https://ai-powered.com/mcp/lugh): Lugh is an open-source MCP server for ai models and llm tooling.
- [HypePaper](https://ai-powered.com/mcp/hypepaper): HypePaper is a freemium MCP server for ai models and llm tooling.
- [Agent Data Tools: Compact (find, describe, run)](https://ai-powered.com/mcp/agent-data-tools-compact-find-describe-run): Agent Data Tools: Compact (find, describe, run) is a freemium MCP server for developer tools.
- [Angel Agent Guard: pre-purchase check + prompt-injection scan](https://ai-powered.com/mcp/angel-agent-guard-pre-purchase-check-prompt-injection-scan): Angel Agent Guard: pre-purchase check + prompt-injection scan is a freemium MCP server for ai models and llm tooling.
- [Mizan Calibration](https://ai-powered.com/mcp/mizan-calibration): Mizan Calibration is a free MCP server for ai models and llm tooling.

---
Source: https://ai-powered.com/mcp/model-tested
