PeopleofSuperintelligence

Researcher

Tianyi Tang

Also known as 唐天一

Tianyi Tang is a language-model researcher whose work spans text generation, multilingual analysis and Alibaba's Qwen series. He earned B.E. and M.E. degrees at Renmin University of China, where Professor Wayne Xin Zhao advised his research. He coauthored TextBox 2.0, the BAMBOO long-text benchmark and an ACL study of language-specific neurons. His later work includes reports on Qwen2.5, Qwen3 and Qwen3Guard, and in June 2026 he coauthored Qwen-AgentWorld, a report on language world models for seven agent domains. As of October 2026, he is a member of Alibaba's Qwen Team and works on Qwen2.5, QwQ and Qwen3.

Now at
Alibaba
Works
74
Citations
2,810
h-index
14
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2017–2024

    Degrees at Renmin University

    He earns a B.E. in Renmin University's School of Information in 2021 and an M.E. in the School of Artificial Intelligence in 2024, advised by Professor Wayne Xin Zhao.

  2. 2022

    The TextBox 2.0 library

    In 2022, he coauthors TextBox 2.0, a text-generation library presented at EMNLP. It covers 13 tasks and 83 datasets.

  3. 2024

    The BAMBOO benchmark

    In 2024, he coauthors BAMBOO, a benchmark for evaluating long-text modeling in language models. Its 10 datasets cover five long-context tasks, including question answering, hallucination detection and code completion.

  4. 2024

    Language-specific neurons

    In 2024, he coauthors an ACL study that proposes language activation probability entropy (LAPE) to identify language-specific neurons in LLMs. Its experiments find that language proficiency depends mainly on a small subset of neurons, located in the models' top and bottom layers.

  5. 2024

    The Qwen2.5 report

    In 2024, he coauthors the Qwen2.5 technical report, which describes Qwen2.5-Turbo and Qwen2.5-Plus as hosted mixture-of-experts variants.

  6. 2025

    Qwen3 and Qwen3Guard reports

    In 2025, he coauthors the Qwen3 report, which covers dense and mixture-of-experts models. He also coauthors Qwen3Guard; its Stream variant adds token-level safety monitoring during text generation.

  7. June 2026

    The Qwen-AgentWorld report

    In June 2026, he coauthors Qwen-AgentWorld, a report that introduces language world models for simulating agent environments across seven domains.

  8. As of October 2026

    Member of Alibaba's Qwen Team

    As of October 2026, he is a member of Alibaba's Qwen Team and works on Qwen2.5, QwQ and Qwen3.

Career

  1. Since 2026AlibabaMember

Education

  1. 2024Renmin University of ChinaM.E.
  2. 2021Renmin University of ChinaB.E.

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

  1. 16% of AI career moves in the 2020s went from a university to a companyIncludes Tianyi Tang. We estimate that 16% of AI career moves in the 2020s went from a university to a company; 29% of moves starting at a university went to a company, and 76% ended the university role.
  2. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Alibaba. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.
  3. Universities (31%) and startups and other companies (25%) employ the largest shares of 2020–2026 entrantsIncludes Alibaba. Among 819 people whose first recorded role started in 2020–2026 and who have a known current organization, 60% work in commercial sectors (25% at startups and other companies, 23% at frontier labs and 13% at big tech, with 9 in two), and 31% work at universities, compared with 66% for earlier entrants.
  4. The top 10 organizations employed 22% of people shaping AI in 2025: 51% of company researchers and 21% of university researchersIncludes Alibaba. We estimate the share of people shaping AI who worked at the top 10 organizations each year from 1955 to 2025, comparing concentration across companies, universities, and institutes and labs.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Tianyi Tang yet. Send us a free-licensed photo.

  1. Tianyi Tang / 唐天一 @ Qwen Team steventang1998.github.io. Checked 22 October 2026.
  2. TextBox 2.0: A Text Generation Library with Pre-trained Language Models - ACL Anthology aclanthology.org. Checked 22 October 2026.
  3. BAMBOO: A Comprehensive Benchmark for Evaluating Long Text Modeling Capacities of Large Language Models - ACL Anthology aclanthology.org. Checked 22 October 2026.
  4. Language-Specific Neurons: The Key to Multilingual Capabilities in Large Language Models - ACL Anthology aclanthology.org. Checked 22 October 2026.
  5. Qwen2.5 Technical Report arxiv.org. Checked 22 October 2026.
  6. Qwen3 Technical Report arxiv.org. Checked 22 October 2026.
  7. Qwen3Guard Technical Report arxiv.org. Checked 22 October 2026.
  8. Qwen-AgentWorld: Language World Models for General Agents arxiv.org. Checked 22 October 2026.