Nonobench
An open-source benchmark of how well LLMs solve nonogram puzzles, from 5x5 to 20x20.
Connect it
Settings → Connectors → Add custom connector
URL: https://www.nonobench.com/mcpSettings → Apps & Connectors → Advanced → Developer mode
Create connector with URL: https://www.nonobench.com/mcp{
"mcpServers": {
"nonobench": {
"url": "https://www.nonobench.com/mcp"
}
}
}claude mcp add --transport http nonobench https://www.nonobench.com/mcp{
"servers": {
"nonobench": {
"type": "http",
"url": "https://www.nonobench.com/mcp"
}
}
}11 tools, verified live
get_leaderboardread-onlyModels ranked by accuracy; defaults to all effort levels for compatibility.
list_providersread-onlyProvider ids, names, families and variant counts.
list_familiesread-onlyModel families, available efforts and best variants.
compare_modelsread-onlySide-by-side core overall and per-size accuracy, cost, latency and token results for model or family names.
get_model_resultsread-onlyAccuracy, cost, latency and token use for one model, broken down by grid size.
list_puzzlesread-onlyThe benchmark puzzles with their ids and row/column clues.
get_puzzleread-onlyOne puzzle, including the clue text models were prompted with. The reference solution is only included on request; some puzzles have several valid solutions.
check_solutionread-onlyCheck a grid against a puzzle's clues, using the same rule as the benchmark grader. Reports which rows and columns do not match.
get_puzzle_resultsread-onlyPer-model outcomes for one puzzle. Answers are omitted unless requested.
get_model_puzzlesread-onlyWhich puzzles one model solved, missed, timed out on, or has not run.
list_runsread-onlyIndividual benchmark runs (one model on one puzzle), optionally with the raw prompt and model output.
Alternatives in Developer Tools
Compare →
ShipStaticA
Deploy static websites, landing pages, and prototypes instantly from AI agents.
RNRelease Notes IndexA
Release notes and changelogs from across the web, indexed for agents. Also via CLI and REST API.
CDclassifier.devA
Sort up to 1,000 texts into your own labels with a calibrated confidence per answer. No API key.
HOHonchoA
Memory that reasons: continual learning for stateful agents. Better context, fewer tokens.
MOMobilerunA
Control real Android and iOS devices with LLM agents — tap, swipe, type, automate flows.
DMDocs MCPA
Remote MCP server for Tandem docs, install guides, SDKs, workflows, and agent setup help.