AIエージェント
SWE-bench_Pro-os
A benchmark testing whether agents can solve long-horizon coding tasks.
★ 539 スターMIT
ガイドを見る →MMLU-Proは、TIGER-AI-Labが公開しているオープンソースのAIエージェントプロジェクトです（GitHub: TIGER-AI-Lab/MMLU-Pro）。GitHubスターは429、フォークは58で、Apache-2.0ライセンスの下で公開されています。GitHubでの最終更新は2026年3月18日です。
The tougher, more robust successor to the MMLU benchmark.
Write down the outcome you want in one sentence.
Clone it locally and check the README.
Get dependencies and tooling in place per the docs.
Follow the minimal example, then adapt it to your needs.
Beyond what MMLU could separate.
Ten options per question raise the bar.
Breadth across knowledge domains.
Code and data, open to all.
Peer-reviewed benchmark design.
Increasingly the test labs must pass.
Download the source and move into the project folder.
git clone https://github.com/TIGER-AI-Lab/MMLU-Pro.git
cd MMLU-ProIsolate dependencies, then install requirements.
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txtLaunch using the entry point shown in the README.
python -m mmlu_pro # check README for the exact commandIt is an open-source project at TIGER-AI-Lab/MMLU-Pro with 58 forks. Check the contributors tab and recent commits to judge how active maintenance is.
Yes — Apache 2.0 allows commercial use and modification. Keep the license and notice files, and note it includes a patent grant.
More answer choices per question and harder items overall — designed to separate frontier models.
Clone the repo and follow the eval instructions; it's code and data, not a service.
Clone the repo and follow the README quickstart — the README is the best starting point.
A benchmark testing whether agents can solve long-horizon coding tasks.
The benchmark that grades AI on real GitHub issues.
A token-conscious MCP server that connects AI tools to Postgres, MySQL, SQLite and more.
A benchmark for tool-agent-user interaction in real-world domains.
An open benchmark for browser AI agents on everyday tasks.
A graduate-level benchmark even Google can't crack.
LibreAIカタログに掲載されたAIエージェントプロジェクトの中で最もスターの多いもの — MMLU-Proが合わなかった場合に検討したい、現実的な候補リストです。
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
カタログの中でMMLU-Proとスター数が近いAIエージェントプロジェクト — 前後の棚もそのまま見て回れます。
A harness-oriented framework for production-grade agent applications.
Safety testing for agents, right in pytest.
Snowflake's Agent World Model: synthesise 1,000 executable MCP tool-use environments for agentic RL.
Agentic memory for LLM agents, from a published research project.
Guardrails for building secure, robust agents.
Benchmark computer-use agents on long, real-world desktop tasks.