Arbor — an autonomous research agent that proposes ideas, edits code, runs experiments, and iterates on a hypothesis tree.
How to install arbor?
ruc-nlpir/arbor
$git clone https://github.com/ruc-nlpir/arborInstalls into the current project.
›Prefer a prompt? Paste this to your agent
Install arbor by running `git clone https://github.com/ruc-nlpir/arbor`, then use it for the current task and follow its documentation at https://github.com/ruc-nlpir/arbor.
README.md
| 1 | <p align="center"> |
| 2 | <img src="assets/hero.svg" alt="Arbor — Optimize anything" width="100%"> |
| 3 | </p> |
| 4 | |
| 5 | |
| 6 | <h1 align="center">🌳 Arbor</h1> |
| 7 | |
| 8 | <h3 align="center">The autonomous research agent that beats Claude Code and Codex by <b>2.5×</b> on the same compute budget</h3> |
| 9 | |
| 10 | <p align="center"> |
| 11 | <a href="https://arxiv.org/pdf/2606.11926"><img src="https://img.shields.io/badge/Paper-arXiv-B31B1B?style=for-the-badge&logo=arxiv&logoColor=white" alt="Paper"></a> |
| 12 | <a href="https://github.com/RUC-NLPIR/Arbor"><img src="https://img.shields.io/badge/Code-GitHub-181717?style=for-the-badge&logo=github&logoColor=white" alt="GitHub"></a> |
| 13 | <a href="https://RUC-NLPIR.github.io/Arbor/"><img src="https://img.shields.io/badge/Project_Page-Live-0E9B9B?style=for-the-badge&logo=githubpages&logoColor=white" alt="Project Page"></a> |
| 14 | <a href="https://RUC-NLPIR.github.io/Arbor/demo.html"><img src="https://img.shields.io/badge/Live_Demo-Watch-FF5C8A?style=for-the-badge&logo=githubpages&logoColor=white" alt="Live Demo"></a> |
| 15 | <a href="https://RUC-NLPIR.github.io/Arbor/docs/"><img src="https://img.shields.io/badge/Docs-Material-526CFE?style=for-the-badge&logo=materialformkdocs&logoColor=white" alt="Docs"></a> |
| 16 | <a href="https://github.com/RUC-NLPIR/Arbor/discussions"><img src="https://img.shields.io/badge/Discussions-Join-5865F2?style=for-the-badge&logo=github&logoColor=white" alt="Discussions"></a> |
| 17 | <a href="LICENSE"><img src="https://img.shields.io/badge/License-Apache_2.0-D22128?style=for-the-badge&logo=apache&logoColor=white" alt="License: Apache 2.0"></a> |
| 18 | </p> |
| 19 | |
| 20 | <p align="center"> |
| 21 | <b>English</b> | <a href="README.zh-CN.md">简体中文</a> |
| 22 | </p> |
| 23 | |
| 24 | <p align="center"> |
| 25 | <i>Give Arbor a benchmark and a goal. It proposes hypotheses, edits code, runs real |
| 26 | experiments, and keeps only the gains that survive held-out data — growing a |
| 27 | <b>hypothesis tree</b> instead of forgetting what failed.</i> |
| 28 | </p> |
| 29 | |
| 30 | > **▶️ Try it in 30 seconds — no API key, no config:** |
| 31 | > |
| 32 | > ```bash |
| 33 | > pip install arbor-agent && arbor replay --demo # watch the hypothesis tree grow live |
| 34 | > ``` |
| 35 | > |
| 36 | > Or **watch it right now in your browser** — nothing to install: **[▶️ Live Demo](https://RUC-NLPIR.github.io/Arbor/demo.html)**. |
| 37 | |
| 38 | ### 🏆 One controller, six tasks — wins the held-out test on all of them |
| 39 | |
| 40 | | Task | Metric | Claude Code | Codex | **Arbor** | |
| 41 | | --- | --- | :---: | :---: | :---: | |
| 42 | | BrowseComp | acc ↑ | 53.33 | 50.00 | **67.67** | |
| 43 | | Terminal-Bench 2.0 | pass ↑ | 71.70 | 73.59 | **77.36** | |
| 44 | | Math-Reasoning Data | gap ↑ | 8.33 | 6.25 | **20.83** | |
| 45 | |
| 46 | Plus **86.36% Any-Medal on MLE-Bench Lite** (GPT-5.5). → [See all six tasks](#-results) · [project page](https://RUC-NLPIR.github.io/Arbor/) · [paper](https://arxiv.org/pdf/2606.11926) · [docs](https://RUC-NLPIR.github.io/Arbor/docs/) |
| 47 | |
| 48 | ## 🎬 Demo |
| 49 | |
| 50 | |
| 51 | https://github.com/user-attachments/assets/49c1a306-d2e9-49d6-9c83-65e38a62df30 |
| 52 | |
| 53 | ## 📣 News |
| 54 | |
| 55 | - **2026-06-30** — **Arbor learns from its own runs.** Each run leaves concrete, reusable findings — a dataset quirk that helped, a trap to avoid; the next similar task recalls them at intake, so the agent starts from experience instead of scratch. 🧠 |
| 56 | - **2026-06-22** — **Built-in literature search & idea novelty checks.** Arbor can now ground its research in prior work via the public [alphaXiv](https://www.alphaxiv.org) API — zero config, no search endpoint or key. Novelty-check any idea before you build it with `arbor idea-check "<your idea>"`, or let the Coordinator vet every new branch automatically. See [Literature Search & Novelty Checks](#-literature-search--novelty-checks). 🔎 |
| 57 | - **2026-06-18** — Arbor was featured by [VentureBeat](https://venturebeat.com/), one of the leading tech media outlets in the US: ["New AI optimization framework beats Claude Code and Codex by 2.5x on the same compute budget"](https://venturebeat.com/orchestration/new-ai-optimization-framework-beats-claude-code-and-codex-by-2-5x-on-the-same-compute-budget). 📰 |
| 58 | - **2026-06-12** — Arbor's native CLI runtime and Agent Skill Suite (Codex / C |
Related
6 picksType
- wikipedia-mcpMCP server for Wikipedia APIMay 2026
- gpt-researcherGPT Researcher is an autonomous agent designed for comprehensive online research on a variety of tasks.Jul 2026
- arxiv-mcp-serverMCP server for accesing arXiv APIAug 2025
- arxiv-mcp-serverA Model Context Protocol server for searching and analyzing arXiv papersJul 2026
- knowledgebase-mcpBioContextAI Knowledgebase MCP server for biomedical agentic AIJan 2026
- mcp-server-serperSerper MCP Server supporting search and webpage scrapingMar 2025
