Binyuan Hui

Research

Hui's early publications centre on text-to-SQL parsing and dialogue, including R²SQL (AAAI 2021) and S²SQL (ACL 2022 Findings) as first author, along with Graphix-T5 and STAR for converting questions into database queries. Hui is a co-first author of BIRD, a benchmark of large, real databases for text-to-SQL systems that was a NeurIPS 2023 spotlight.¹⁸

Hui's homepage lists research interests in foundation models (pre-training and post-training), reasoning models for code, coding agents and computer-use agents, and states a goal of building autonomous coding models.¹

Open models and agents

Hui was part of the Qwen team at Alibaba Group, which released its first Qwen model in April 2023, and is an author of the Qwen, Qwen2 and Qwen3 technical reports and first author of the Qwen2.5-Coder report. In July 2023 Hui joined the BigCode collaboration, contributing to OctoPack and StarCoder 2.¹⁶⁹⁷⁵⁴

Hui is described as the initiator of OpenDevin, the open platform for software-development agents later renamed OpenHands and released under the MIT license; the 2024 OpenHands paper lists Hui among its authors.¹³¹⁰²

Work and contributions

• OpenHands (formerly OpenDevin) Open-source code

  Described on Hui's homepage as the initiator of OpenDevin, and a co-author of "OpenHands: An Open Platform for AI Software Developers as Generalist Agents" (2024), which introduces OpenHands (f.k.a. OpenDevin) and notes its release under the MIT license. The repository's commit history lists commits by Hui's account from March to May 2024, including support for the BIRD benchmark and early front-end and repository fixes, and Hui's GitHub profile pins the OpenHands repository.^1311102

• StarCoder2 Open weights

  Co-author of "StarCoder 2 and The Stack v2: The Next Generation" (2024) from the BigCode project, which introduces the StarCoder2 family of code models and makes the weights available under an OpenRAIL license.⁴¹

• Qwen2.5-Coder Published research

  First author of the Qwen2.5-Coder Technical Report (2024); also a co-author of the Qwen Technical Report (2023) and the Qwen3 Technical Report (2025).⁵⁶⁷

• BIRD text-to-SQL benchmark Published research

  Co-first author of "Can LLM Already Serve as A Database Interface? A BIg Bench for Large-Scale Database Grounded Text-to-SQLs" (BIRD), a NeurIPS 2023 spotlight paper on text-to-SQL evaluation.⁸¹

Timeline

1. 2021Published R²SQL, on context-dependent semantic parsing, as first author at AAAI 2021.¹
2. Apr 2023The Qwen team at Alibaba Group released Qwen, an open large language model.¹
3. Jul 2023Joined the BigCode open scientific collaboration on code models.¹
4. Sep 2023The BIRD text-to-SQL benchmark was accepted at NeurIPS 2023 as a spotlight paper.¹⁸
5. Feb 2024Co-authored "StarCoder 2 and The Stack v2 - The Next Generation".⁴
6. Jul 2024Co-authored the Qwen2 Technical Report and the OpenHands platform paper.⁹³
7. Sep 2024Released the Qwen2.5-Coder Technical Report as first author.⁵
8. May 2025Co-authored the Qwen3 Technical Report.⁷

Documented roles

The sources do not confirm a current role; these are the roles they document.

• Staff Research Scientist, Qwen Team, Alibaba Group (Qwen Team)¹²

Honors and recognition

• WAIC Yunfan Award (Rising Stars) · World Artificial Intelligence Conference, 2022¹

Related people

• Harrison ChaseAgents · Developer tools
• Michele CatastaAgents · Language models
• Omar KhattabDeveloper tools · Language models
• Graham NeubigAgents · Language models
• Robert BrennanAgents · Developer tools
• Peter SteinbergerAgents · Developer tools
• Boris ChernyDeveloper tools · Agents
• Shunyu YaoAgents · Language models
• Kanjun QiuAgents · Developer tools
• John YangAgents · Developer tools