Nonobench MCP Server

by mauricekleine Your server? Claim it

answering

Nonobench is answering right now. Last checked 9 min ago. It exposes 11 tools. Last commit 31 Oct 2026.

An open-source benchmark of how well LLMs solve nonogram puzzles, from 5x5 to 20x20.

Uptime history 34 days of history · worst day 99%

34 days agonow

100.0%

Uptime 24h

91 of 91 checks

Tools

read from the server

259 ms

Response time

average over 24h

Stars

last commit 31 Oct 2026

Nothing serious here today

Today is the operative word: we check Nonobench every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.

Connect this server

Endpoint below is the one we actually reach during checks — not the one copied from a README. Last verified 9 min ago.

Claude Code Claude Desktop Codex Cursor VS Code

run in your terminal

claude mcp add nonobench --transport http https://www.nonobench.com/mcp

Settings → Connectors → Add custom connector, then paste this address

https://www.nonobench.com/mcp

~/.codex/config.toml

[mcp_servers.nonobench]
url = "https://www.nonobench.com/mcp"

.cursor/mcp.json

{
  "mcpServers": {
    "nonobench": {
      "url": "https://www.nonobench.com/mcp"
    }
  }
}

.vscode/mcp.json

{
  "servers": {
    "nonobench": {
      "type": "http",
      "url": "https://www.nonobench.com/mcp"
    }
  }
}

Available tools 11

Read directly from the server with tools/list, grouped by what they act on. If a tool disappears, we record the date.

model

get_model_puzzles

Which puzzles one model solved, missed, timed out on, or has not run.

get_model_results

Accuracy, cost, latency and token use for one model, broken down by grid size.

puzzle

get_puzzle

One puzzle, including the clue text models were prompted with. The reference solution is only included on request; some puzzles have several valid solutions.

get_puzzle_results

Per-model outcomes for one puzzle. Answers are omitted unless requested.

compare

compare_models

Side-by-side core overall and per-size accuracy, cost, latency and token results for model or family names.

families

list_families

Model families, available efforts and best variants.

leaderboard

get_leaderboard

Models ranked by accuracy; defaults to all effort levels for compatibility.

providers

list_providers

Provider ids, names, families and variant counts.

puzzles

list_puzzles

The benchmark puzzles with their ids and row/column clues.

runs

list_runs

Individual benchmark runs (one model on one puzzle), optionally with the raw prompt and model output.

solution

check_solution

Check a grid against a puzzle's clues, using the same rule as the benchmark grader. Reports which rows and columns do not match.

Endpoints

URL	Transport	State	Latency	Checked
https://www.nonobench.com/mcp	streamable-http	answering	199 ms	9 min ago

Alternatives to Nonobench

same job, measured the same way

Nonogram MCP

by mrfentmen

Solve, check and render nonogram (picross) puzzles from row and column clues - fully local.

469 installs/wk local only

SearchForge

by divyanshu-iitian

Free open-source web search and page reading for LLMs, AI agents, and RAG.

local only

The Aggregate — LLM benchmark aggregate

by theaggregate

LLM rankings from public benchmarks, model comparisons and source-linked results. Updated daily.

8 tools answering

Exa Search

by pollinations

Search the live web and fetch clean content from source pages.

answering

All SVG Icons

by allsvgicons

Search and fetch SVG icons from 250,000+ icons across 200+ open-source icon sets.

answering

Levels.fyi

by levels

Salary, level and job data from 1M+ pay reports. Benchmark pay, evaluate offers, search open jobs.

30 tools answering

Gondar

by fluxonlab

Search open grants, tenders and fellowships from official sources, with facts and official links.

11 tools answering

Sauce API MCP

by saucelabs

An open-source MCP server that provides LLM access to the Sauce Labs API

194 installs/wk local only

Nonobench — questions

Answers built from our own checks of this server.

What can Nonobench do?

It exposes 11 tools, read directly from the server on our last check. Among them: check_solution, compare_models, get_leaderboard, get_model_puzzles, get_model_results, get_puzzle and 5 more. The full list with descriptions is on this page — we take it from the server itself via tools/list, not from a README. How MCP servers expose tools in the first place →

What is Nonobench mostly used for?

Its tools cluster around model and puzzle. That is what this server is built to work with — the grouping comes from the actual tool names, not from a category we assigned.

Is Nonobench working right now?

We send a real MCP handshake every 15 minutes. Over the last 24 hours 91 of 91 checks got a reply (100.0%), average response time 259 ms. The bar chart above shows every period we have measured.

How do I connect Nonobench?

Copy the ready config from this page. We generate it for Claude Code, Claude Desktop, Codex, Cursor and VS Code, each in the form that client actually reads. It is a remote server, so there is nothing to install — the client connects to the address.

Does Nonobench need an API key?

No. Nonobench completed a full MCP handshake with us as an anonymous client and listed its tools without asking for anything. All 11 of them are readable on this page. This is what we observed, not what the docs claim.

How fast is Nonobench?

It answers our handshake in 259 ms on average, which is faster than 55% of all working MCP servers we measure. The comparison comes from our own checks across the whole registry, every 15 minutes.

Is Nonobench open source?

The code is open source: it is published under the MIT licence, written in TypeScript and 10 stars on GitHub. The licence covers the code in the repository, not the hosted service at nonobench.com, which runs on its operator's own terms.