PeopleofSuperintelligence

Researcher

Beichen Zhang

Beichen Zhang is an AI researcher whose work spans language models and mathematical reasoning. He co-authored ChatCoT in 2023, a framework that brings tools into multi-turn chain-of-thought reasoning. His later papers include Qwen2.5-Math, which develops math-specific models through self-improvement; START, a tool-integrated reasoner; and ProcessBench, a benchmark for errors in mathematical reasoning. He was a core contributor to Qwen3 and co-authored a 2026 survey of large language models. As of October 2026, he is a research scientist in Tongyi Lab.

Now at
Tongyi Lab
Works
18
Citations
542
h-index
6

Publication numbers from OpenAlex, as of 18 October 2026.

Timeline

  1. 2023

    ChatCoT and tool use

    In 2023, he coauthors ChatCoT, a tool-augmented chain-of-thought framework that treats reasoning as multi-turn conversation.

  2. 2024

    Building Qwen2.5-Math

    In 2024, he coauthors Qwen2.5-Math, a series of math-specific language models built through self-improvement across pre-training, post-training, and inference.

  3. 2025

    START with tools

    In 2025, he coauthors START, a tool-integrated long chain-of-thought model that uses code execution for computation, self-checking, and self-debugging.

  4. 2025

    Core contributor to Qwen3

    In 2025, he is a core contributor to the Qwen3 technical report. The Qwen team at Alibaba Cloud describes Qwen3 as its large language model series.

  5. 2025

    Finding errors in math reasoning

    In 2025, he coauthors ProcessBench, a benchmark that asks models to locate the earliest erroneous step in mathematical reasoning.

  6. 2026

    Surveying large language models

    In 2026, he coauthors A Survey of Large Language Models, which reviews pre-training, post-training, use, and evaluation.

  7. As of October 2026

    Research scientist in Tongyi Lab

    As of October 2026, he is a research scientist in Tongyi Lab.

Career

  1. NowTongyi LabResearch scientist
  2. 2025–2026Beijing Institute of Technology
  3. 2014–2020National University of Defense Technology

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

Most-cited works

  1. Overview of the key technologies of combined cycle engine precooling systems and the advanced applications of micro-channel heat transfer2014 · Aerospace Science and Technology · 155 citations
  2. Qwen3 Technical Report2025 · 147 citationsWith Baosong Yang, Fei Huang, Le Yu, Rui Men and Zihan Qiu
  3. Qwen2.5 Technical Report2024 · 122 citationsWith Baosong Yang, Fei Huang, Le Yu, Rui Men and Zihan Qiu
  4. A high-frequency artificial nerve based on homogeneously integrated organic electrochemical transistors2025 · Nature Electronics · 62 citations
  5. Experimental investigation of nitrogen flow boiling heat transfer in a single mini-channel2020 · Journal of Zhejiang University. Science A · 32 citations

From OpenAlex, as of 18 October 2026.

  1. The company share of AI researchers’ top papers rose from 3% to 32%Includes National University of Defense Technology. We estimate where leading AI researchers worked when their most-cited papers appeared. The company share rose from 3% in 1975–1979 to 32% in 2020–2024.
  2. Stanford University trained or employed 27 of the 111 AI major-prize winners, more than any other universityIncludes Beijing Institute of Technology. Stanford University trained or employed 27 of the 111 AI major-prize winners, more than any other university. We compare doctorate and employment links for major-prize and best-paper winners.
  3. Google had the most people shaping AI every year from 2014 to 2025Our estimates of which organizations employed the most people shaping AI from 1961 to 2025, with a year-by-year chart.
  4. The United States employs an estimated 53% of people in our data with a known employment countryOur estimates of where AI careers are concentrated now and how countries’ shares of dated positions changed over time.

Colleagues

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

Sources and credits

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

  1. ChatCoT: Tool-Augmented Chain-of-Thought Reasoning on Chat-based Large Language Models arxiv.org. Checked 18 October 2026.
  2. Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvement arxiv.org. Checked 18 October 2026.
  3. START: Self-taught Reasoner with Tools arxiv.org. Checked 18 October 2026.
  4. Qwen3 Technical Report arxiv.org. Checked 18 October 2026.
  5. Qwen3 | QwenLM on GitHub github.com. Checked 18 October 2026.
  6. ProcessBench: Identifying Process Errors in Mathematical Reasoning arxiv.org. Checked 18 October 2026.
  7. A Survey of Large Language Models journal.hep.com.cn. Checked 18 October 2026.
  8. Qwen3 Technical Report | Report Profile | LLMpeople llm-people.com. Checked 18 October 2026.