# China AI Atlas — Site Facts

Generated from the atlas's own dataset (network.json + person-profiles.json + lab-financials.json + market-caps.json + model-arena.json).
Use this as a primary, structured fact source when writing the launch report.
All names follow the "English (中文)" pairing used on the site.

## DeepSeek (深度求索)

Hangzhou foundation-model lab spun off from quant fund High-Flyer (幻方量化). R1 made DeepSeek the cost-curve reference point for open reasoning models. The Information reported $70.7M of revenue for January-July 2026 and a $400M-$500M annualized API-revenue pace.

- **Founded:** 2023-07
- **HQ:** Hangzhou (杭州)
- **Parent / spun out of:** High-Flyer (幻方量化)
- **Valuation note:** Reported $45-50B first external raise talks in May 2026; reported target raise size $3-4B. Self-funded via High-Flyer through 2025.
- **Founders:** Liang Wenfeng (梁文锋)
- **Key people:** Deli Chen (陈德立) — Researcher, DeepSeek; Yu Wu (吴俣) — Head of LLM Alignment Team, DeepSeek; Zhihong Shao (邵智宏) — Research Scientist, DeepSeek; Zhibin Gou (苟志斌) — Core researcher, DeepSeek; Dejian Yang (杨德建) — Researcher, DeepSeek; Junxiao Song (宋俊潇) — Principal Researcher, DeepSeek; Damai Dai (代达劢) — Deep Learning Researcher, DeepSeek; Qihao Zhu (朱启昊) — Researcher, DeepSeek
- **Notable alumni / departures:** Daya Guo (郭达雅); Luo Fuli (罗福莉)
- **Investors:** China Big Fund (国家集成电路产业投资基金); Huajin Capital (华金资本); Zhejiang Oriental (浙江东方)
- **Government backing:** Hangzhou Municipal Government (杭州市政府); Zhejiang Provincial Government (浙江省政府)
- **Notable models:** DeepSeek-Coder, DeepSeek-V2, DeepSeek-V3, DeepSeek-R1, DeepSeek-V3.2
- **REVENUE:** $70.7M (Jan-Jul 2026) (rumored, as of 2026-07-31, The Information, https://www.theinformation.com/articles/deepseeks-revenue-reaches-70-million-july-tenfold-jump-2025)
- **ARR:** $400M-$500M reported run-rate (rumored, as of 2026-07-15, The Information, https://www.theinformation.com/articles/deepseeks-annualized-revenue-nears-500-million-boosting-fundraise-ipo-plans)
- **Model arena (OpenRouter pricing; AAI = Artificial Analysis Index; as of 2026-09-03):** DeepSeek V4 Pro (AAI 45.3) $1.034/$2.069 per 1M in/out; DeepSeek V4 Flash (AAI 42.1) $0.081/$0.162 per 1M in/out
- **Site:** https://www.deepseek.com

## Zhipu AI / Z.ai (智谱AI)

Spun out of Tsinghua's Knowledge Engineering Group (KEG); founded by professors Tang Jie (唐杰) and Li Juanzi (李娟子); CEO Zhang Peng (张鹏).

- **Founded:** 2019-06
- **HQ:** Beijing (北京)
- **Parent / spun out of:** Tsinghua University (清华大学)
- **Founders:** Tang Jie (唐杰); Li Juanzi (李娟子)
- **Key people:** Zhang Peng (张鹏) — CEO, Zhipu; Yuxiao Dong (董昱潇) — Tsinghua KEG Associate Professor; senior collaborator on GLM models; Zhengxiao Du (杜正晓) — Tech Lead, Zhipu (since January 2023)
- **Notable alumni / departures:** Ming Ding (丁铭)
- **Investors:** Alibaba Group (阿里巴巴); Tencent (腾讯); HongShan (formerly Sequoia China) (红杉中国); Xiaomi (小米); Meituan (美团); Ant Group (蚂蚁集团); Prosperity7 (Saudi Aramco); Pudong Venture Capital (浦东创投); Zhangjiang Group (张江集团)
- **Government backing:** Beijing Municipal Government (北京市政府); Hangzhou Municipal Government (杭州市政府)
- **Notable models:** GLM-130B, ChatGLM, GLM-4, GLM-5.2, CogVLM, CogVideoX
- **Market cap:** $64.2B (2513.HK, HKD, as of 2026-09-04, Yahoo Finance); 981 employees
- **REVENUE:** RMB954M (~$132M), 1H2026 (confirmed, as of 2026-06-30, Z.ai HKEX interim results, https://www1.hkexnews.hk/search/titlesearch.xhtml?category=0&lang=EN&market=SEHK&stockId=1000286748)
- **ARR:** $1.6B monthly annualized run-rate (confirmed, as of 2026-08-31, Z.ai results reporting / Zhitong, https://news.futunn.com/post/78565669/zhipu-02513-maas-arr-has-reached-1-6-billion-representing)
- **HEADCOUNT:** 1,094 (confirmed, as of 2026-05-28, Yahoo Finance, https://finance.yahoo.com/quote/2513.HK/)
- **Model arena (OpenRouter pricing; AAI = Artificial Analysis Index; as of 2026-09-03):** GLM-5.3 (AAI 59.5) $1.4/$4.4 per 1M in/out
- **Site:** https://www.zhipuai.cn

## Moonshot AI / Kimi (月之暗面)

Founded by ex-Tsinghua KEG / CMU PhD Yang Zhilin (杨植麟). Long-context flagship Kimi consumer chatbot.

- **Founded:** 2023-03
- **HQ:** Beijing (北京)
- **Valuation note:** $2B reported round at a $20B valuation in May 2026.
- **Founders:** Yang Zhilin (杨植麟); Zhang Yutao (张宇韬); Zhou Xinyu (周昕宇); Wu Yuxin (吴育昕)
- **Key people:** Su Jianlin (苏剑林) — Researcher, Moonshot AI
- **Investors:** Alibaba Group (阿里巴巴); Tencent (腾讯); HongShan (formerly Sequoia China) (红杉中国); IDG Capital (IDG资本); Gaorong Capital (高榕资本)
- **Government backing:** Beijing Municipal Government (北京市政府)
- **Notable models:** Kimi K1.5, Kimi K2, Kimi K2.5, Kimi K2.6
- **ARR:** ~$300M reported ARR (rumored, as of 2026-06-30, Bloomberg, https://www.bloomberg.com/news/articles/2026-07-19/china-s-moonshot-plans-ipo-in-six-months-after-ai-breakthrough)
- **Model arena (OpenRouter pricing; AAI = Artificial Analysis Index; as of 2026-09-03):** Kimi K3 (AAI 59.7) $3/$15 per 1M in/out; Kimi K2.6 (AAI 45.1) $0.95/$4 per 1M in/out
- **Site:** https://www.moonshot.cn

## MiniMax (稀宇科技)

Shanghai foundation-model lab founded by ex-SenseTime executives.

- **Founded:** 2021-12
- **HQ:** Shanghai (上海)
- **Founders:** Yan Junjie (闫俊杰); Zhou Yucong (周彧聪)
- **Key people:** Yun Yeyi (贠烨祎) — Executive Director and COO, MiniMax; Zhao Pengyu (赵鹏宇) — Executive Director and LLM Research Leader, MiniMax (June 2025)
- **Notable alumni / departures:** Yang Bin (杨斌); Yongyi Hu (胡咏毅)
- **Investors:** Alibaba Group (阿里巴巴); HongShan (formerly Sequoia China) (红杉中国); Hillhouse Capital (高瓴资本); miHoYo (米哈游)
- **Government backing:** Shanghai SASAC / Municipal Gov (上海市国资委)
- **Notable models:** abab series, MiniMax-01 (Lightning Attention), M1 / M2.5 / M2.7, Hailuo (video), Talkie, Xingye
- **Market cap:** $16.2B (0100.HK, HKD, as of 2026-09-04, Yahoo Finance); 568 employees
- **REVENUE:** $116.6M (1H2026) (confirmed, as of 2026-06-30, MiniMax H1 results, https://www.minimax.io/news/minimax-announces-first-half-2026-financial-results-1787744160)
- **ARR:** $800M+ company-disclosed ARR (confirmed, as of 2026-08-26, Yicai / MiniMax earnings call, https://www.yicai.com/news/103335505.html)
- **HEADCOUNT:** 415 (confirmed, as of 2026-05-28, Yahoo Finance, https://finance.yahoo.com/quote/0100.HK/)
- **Model arena (OpenRouter pricing; AAI = Artificial Analysis Index; as of 2026-09-03):** MiniMax M3 (AAI 45.4) $0.3/$1.2 per 1M in/out
- **Site:** https://www.minimaxi.com

## 01.AI (零一万物)

Founded by Kai-Fu Lee (李开复), incubated by Sinovation Ventures. Exited frontier pre-training Q1 2025.

- **Founded:** 2023-03
- **HQ:** Beijing (北京)
- **Valuation note:** Reached unicorn status ($1B+) November 2023. ~$300M total raised. Pre-training and infra teams absorbed into Alibaba Tongyi/Qwen, Q1 2025.
- **Founders:** Kai-Fu Lee (李开复)
- **Notable alumni / departures:** Wenhao Huang (黄文灏); Guoyin Wang (王国胤); Zonghong Dai (Ethan Dai) (戴宗宏)
- **Investors:** Alibaba Group (阿里巴巴); Sinovation Ventures (创新工场); Xiaomi (小米)
- **Notable models:** Yi-34B, Yi-VL, Yi-Lightning, Wanzhi
- **Site:** https://www.01.ai

## Baichuan AI (百川智能)

Founded by ex-Sogou CEO Wang Xiaochuan (王小川). Pivoted to healthcare AI mid-2024; flagship is 'Futang Baichuan' pediatric AI model with Beijing Children's Hospital.

- **Founded:** 2023-04
- **HQ:** Beijing (北京)
- **Valuation note:** ~$2.8B last disclosed valuation / ~$690M Series A in 2024; now pivoted to healthcare AI.
- **Founders:** Wang Xiaochuan (王小川); Ru Liyun (茹立云)
- **Notable alumni / departures:** Hong Tao (洪涛); Chen Weipeng (陈未鹏); Wang Bingning (王炳宁)
- **Investors:** Alibaba Group (阿里巴巴); Tencent (腾讯); Xiaomi (小米); Beijing AI Industry Investment Fund (北京人工智能产业基金); Shenzhen Capital Group (深创投)
- **Government backing:** Beijing Municipal Government (北京市政府)
- **Notable models:** Baichuan-7B/13B, Baichuan2, Baichuan3, Baichuan-Omni, Baichuan-M1 (medical)
- **Site:** https://www.baichuan-ai.com

## StepFun (阶跃星辰)

Shanghai foundation-model lab founded by ex-MSRA Chief Scientist Jiang Daxin (姜大昕). Densest MSRA-diaspora cluster. Yin Qi (印奇, Megvii co-founder) became chairman January 2026.

- **Founded:** 2023-04
- **HQ:** Shanghai (上海)
- **Valuation note:** Reported nearly $2.5B new financing in May 2026 and a planned Hong Kong IPO at around a $10B valuation.
- **Founders:** Jiang Daxin (姜大昕); Jiao Binxing (焦斌星); Zhu Yibo (朱亦博)
- **Key people:** Yin Qi (印奇) — Chairman, StepFun (January 2026); Chairman, Megvii; Chairman, Qianli Tech (Geely); Gang Yu (余刚) — Principal Researcher AI Lab, StepFun (since 2024)
- **Investors:** Tencent (腾讯); Qiming Venture Partners (启明创投); 5Y Capital (五源资本); Pudong Venture Capital (浦东创投); Shanghai SSCIC / Shanghai Guotou (Fortera) (上海国投 / 上海国际集团); China Life Private Equity (国寿资本)
- **Government backing:** Shanghai SASAC / Municipal Gov (上海市国资委); Pudong New Area (浦东新区)
- **Notable models:** Step-1V, Step-2 (1T MoE), Step-Video-T2V, Step-Audio, Step-3 (321B MoE)
- **REVENUE:** ~RMB500M (~$69M) (confirmed, as of 2026-04-15, CMBI citing Caijing, https://www.cmbi.com.hk/upload/202604/20260415324700.pdf)
- **Model arena (OpenRouter pricing; AAI = Artificial Analysis Index; as of 2026-09-03):** Step 3.7 Flash (AAI 30.9) $0.2/$1.15 per 1M in/out
- **Site:** https://www.stepfun.com

## Qwen team @ Alibaba (通义千问 / 通义实验室)

Tongyi Lab established end of 2022 by merging DAMO language/vision/AI teams into Alibaba Cloud. 36Kr and TMTPost frame Qwen as Alibaba's model layer for cloud AI revenue, Bailian MaaS, open-source distribution, and agent/coding benchmarks; Qwen itself has no standalone valuation or P&L.

- **Founded:** 2022
- **HQ:** Hangzhou (杭州)
- **Parent / spun out of:** Alibaba Group (阿里巴巴)
- **Key people:** Guoyin Wang (王国胤) — (former) Co-founder & Head of Post-Training, 01.AI — now leading Alibaba Tongyi Qwen Pilot group; Zhou Jingren (周靖人) — CTO Alibaba Cloud Intelligence; head of Tongyi Lab; Deputy Head of DAMO; Chief AI Architect; Fei Huang (黄飞) — Chief Scientist & Senior Director, Language Technologies Lab, DAMO Academy; Hao Zhou (周浩) — New Qwen tech lead (March 2026), replacing Lin Junyang; Shuai Bai (白帅) — Lead author Qwen3-VL; head of Qwen vision-language line; An Yang — Lead author Qwen3 Technical Report; pre-training lead
- **Notable alumni / departures:** Lin Junyang (林俊旸); Bowen Yu (郁博文); Zhou Chang (周畅); Luo Fuli (罗福莉); Binyuan Hui (惠彬原)
- **Notable models:** Qwen, Qwen2.5, Qwen3, Qwen3-VL, Qwen3.7-Max (reported), Tongyi Qianwen
- **Model arena (OpenRouter pricing; AAI = Artificial Analysis Index; as of 2026-09-03):** Qwen3.6 Plus (AAI 40.5) $0.325/$1.95 per 1M in/out
- **Site:** https://qwen.ai

## ByteDance Doubao + Seed (豆包 / Seed)

ByteDance's Doubao/Seed foundation-model org (Beijing + Bay Area). Wu Yonghui (吴永辉, ex-Google DeepMind VP) joined as Seed lead February 2025; recent coverage emphasizes Doubao's consumer feedback loop, Seedance video, Volcano Engine token share, and ByteDance's experiments around phone/on-device agent access.

- **Founded:** 2023
- **HQ:** Beijing (北京)
- **Parent / spun out of:** ByteDance (字节跳动)
- **Key people:** Daya Guo (郭达雅) — Researcher, DeepSeek (reportedly departed for ByteDance Seed early 2026); Ming Ding (丁铭) — Now at ByteDance Seed (formerly Zhipu); Wenhao Huang (黄文灏) — ByteDance large-model team; former 01.AI technical co-founder / VP Algorithms; Bowen Yu (郁博文) — Now at ByteDance Seed (joined March 12, 2026 as post-training lead, Visual Model & Multimodal Interaction Team); Wu Yonghui (吴永辉) — Head of Seed foundational research, ByteDance (since February 2025); Zhu Wenjia (朱文佳) — Leads Doubao large-model applications, ByteDance; Wang Mingxuan (王明轩) — Senior Researcher, ByteDance Seed; Co-Director SIA Lab (joint w/ Tsinghua AIR); Zhou Chang (周畅) — Multimodal Interaction & World-Models lead, ByteDance Seed (joined September 2024)
- **Notable alumni / departures:** Zhu Yibo (朱亦博); Yang Jianchao (杨建超); Jiashi Feng (冯佳时); Qiao Mu (乔木)
- **Notable models:** Doubao 1.5 Pro, Doubao 2.0, Seedance 2.0 (video), Doubao-Seed-Code, Seed1.5-VL
- **Site:** https://seed.bytedance.com

## Hunyuan team @ Tencent (混元)

Tencent's foundation-model org. Yao Shunyu (姚顺雨, 28, ex-OpenAI) appointed Chief AI Scientist December 2025. Tencent AI Lab dissolved into Hunyuan March 2026. Recruited Microsoft's WizardLM team in 2025.

- **Founded:** 2023-09
- **HQ:** Shenzhen (深圳)
- **Parent / spun out of:** Tencent (腾讯)
- **Key people:** Yao Shunyu (姚顺雨) — Chief AI Scientist, Tencent (December 2025); Lu Shan (卢山) — President, Tencent TEG (Technology and Engineering Group); Liu Yuhong (刘煜宏) — VP Tencent Cloud; Head of AI Data Department, Hunyuan; Han Hu (胡瀚) — Head of Multimodal Large Model R&D, Tencent Hunyuan (joined 2024); WANG Di (王迪) — Deputy GM, LLM Department, Tencent Hunyuan; CAN Xu (徐灿) — Senior Researcher, Tencent Hunyuan (joined May 2025); Qingfeng Sun (孙庆峰) — Senior Researcher, Tencent Hunyuan (joined May 2025); Weijie Kong — Tencent Hunyuan video generation tech lead; SUN Xingwu (孙兴武) — Tencent Hunyuan LLM department researcher
- **Notable models:** Hunyuan, Hunyuan-Large, Hunyuan Turbo S, Hunyuan T1, HunyuanVideo, Hunyuan3D
- **Model arena (OpenRouter pricing; AAI = Artificial Analysis Index; as of 2026-09-03):** Hunyuan A13B $0.14/$0.57 per 1M in/out
- **Site:** https://hunyuan.tencent.com

## Xiaomi MiMo (小米 MiMo)

Xiaomi's foundation-model team within the AI Lab. Established April 2023; rebranded as 'MiMo' around its 2025 open-source release. Luo Fuli (罗福莉, ex-DeepSeek-V2 core developer) became team head November 2025.

- **Founded:** 2023-04
- **HQ:** Beijing (北京)
- **Parent / spun out of:** Xiaomi (小米)
- **Key people:** Luo Fuli (罗福莉) — Head of Xiaomi MiMo large-model team (November 2025); Wang Bin (王斌) — Director, Xiaomi AI Lab; Vice Chairman of Xiaomi Technical Committee; Bingquan Xia (夏炳权) — Tech lead, Xiaomi MiMo LLM-Core
- **Notable alumni / departures:** Luan Jian (栾剑)
- **Notable models:** MiLM, MiMo-7B, MiMo-7B-RL, MiMo-VL-7B-RL, MiMo-V2 (Flash MoE / Pro / Omni), MiMo-V2.5
- **Model arena (OpenRouter pricing; AAI = Artificial Analysis Index; as of 2026-09-03):** MiMo V2.5 Pro (AAI 42.9) $0.435/$0.87 per 1M in/out
- **Site:** https://github.com/XiaomiMiMo

## Meituan LongCat (美团龙猫)

Meituan-owned LongCat large-model family. LongCat-Flash-Chat was released and open-sourced September 1, 2025 as a 560B MoE model that activates roughly 18.6B-31.3B parameters per token and emphasizes efficient inference, agent/tool-use, and local-services workflows.

- **Founded:** 2025-09
- **HQ:** Beijing (北京)
- **Parent / spun out of:** Meituan (美团)
- **Key people:** Wang Jingang (王金刚) — Technology Director, Meituan LongCat large-model team; Pan Xin (潘欣) — Multimodal AI Innovation Lead, Meituan (joined December 2025); Jianing Wang (王嘉宁) — Researcher, Meituan LongCat; first author, LongCat-Flash-Prover; Xunliang Cai (蔡勋梁) — Researcher, Meituan LongCat video direction
- **Notable models:** LongCat-Flash-Chat, LongCat-Flash-Thinking, LongCat-Flash-Omni, LongCat-Image, LongCat-Video, LongCat-Next
- **Site:** https://longcat.ai

## High-Flyer (幻方量化)

Quant hedge fund co-founded by Liang Wenfeng. AI division established 2019; spun off DeepSeek in 2023.

- **Founded:** 2016
- **HQ:** Hangzhou (杭州)
- **Founders:** Liang Wenfeng (梁文锋)

## Megvii (Face++) (旷视科技)

Computer vision company founded by Tsinghua Yao Class trio (Yin Qi, Tang Wenbin, Yang Mu). Important predecessor lab — Yin Qi went on to chair StepFun.

- **Founded:** 2011
- **HQ:** Beijing (北京)
- **Founders:** Yin Qi (印奇)
- **Notable alumni / departures:** Zhou Xinyu (周昕宇); Gang Yu (余刚)

## Sogou (搜狗)

Search engine spun off from Sohu by Wang Xiaochuan. Acquired by Tencent 2021. Wang plus 3 ex-Sogou execs founded Baichuan in 2023.

- **Founded:** 2004
- **HQ:** Beijing (北京)
- **Founders:** Wang Xiaochuan (王小川)
- **Notable alumni / departures:** Ru Liyun (茹立云); Hong Tao (洪涛); Chen Weipeng (陈未鹏); Wang Bingning (王炳宁)
- **Investors:** Tencent (腾讯)

## SenseTime (商汤科技)

Computer vision unicorn. SenseTime alumni cluster founded MiniMax (Yan Junjie + Zhou Yucong + Yun Yeyi).

- **Founded:** 2014
- **HQ:** Shanghai (上海)
- **Notable alumni / departures:** Yan Junjie (闫俊杰); Zhou Yucong (周彧聪); Yun Yeyi (贠烨祎)

## Recurrent AI (循环智能)

Yang Zhilin and Zhang Yutao's first venture before Moonshot.

- **Founded:** 2016
- **HQ:** Beijing (北京)
- **Founders:** Yang Zhilin (杨植麟); Zhang Yutao (张宇韬)

## Tsinghua KEG (Knowledge Engineering Group) (清华大学知识工程实验室)

Tsinghua CS lab led by Tang Jie + Li Juanzi. Direct origin of Zhipu (2019) and Yang Zhilin's training (Moonshot).

- **Founded:** n/a
- **HQ:** Beijing (北京)

## Tencent AI Lab (腾讯AI Lab)

Tencent's research lab; led by Zhang Tong for years. Dissolved March 2026 and merged into Hunyuan team.

- **Founded:** 2016
- **HQ:** Shenzhen (深圳)


# Researchers

### Deli Chen (陈德立)
Researcher, DeepSeek

Deli Chen (陈德立) is a researcher at DeepSeek working on language models, alignment mechanisms, training strategies, and model generalization. He holds a bachelor's (Information Management & Information Systems) and a master's (EECS, 2020) from Peking University, interned at Mizuho Securities (Tokyo) and Tencent WeChat AI, and was a researcher at Tencent WeChat AI (2021–2023) before joining DeepSeek in 2023. He is a contributor to the DeepSeek V2, V3, and R1 releases, with 13,000+ Google Scholar citations. He represented DeepSeek on the “Hangzhou Six Dragons” panel at the 2025 World Internet Conference in Wuzhen.
- Known for: DeepSeek V2 / V3 / R1 contributor; Alignment, training & generalization research
- Education: Peking University (北京大学)

### Liang Wenfeng (梁文锋)
Founder & CEO, DeepSeek; co-founder & CEO, High-Flyer

Liang Wenfeng (梁文锋, born 1985) is the founder and CEO of DeepSeek, the Hangzhou-based AI research lab that gained global recognition with its DeepSeek-R1 and DeepSeek-V3 models. Before founding DeepSeek in 2023, Liang co-founded High-Flyer Capital Management (幻方量化), a quantitative trading firm, in 2015. High-Flyer grew to manage over 100 billion yuan (~$14B) in assets by 2021, becoming one of China's largest quant funds. Liang holds a Bachelor's and Master's degree in Information and Communication Engineering from Zhejiang University (graduated 2010). In January 2025, DeepSeek's R1 model reportedly rivaled OpenAI's ChatGPT at a fraction of the training cost, making global headlines. Liang was named among Nature's top 10 'people who shaped science in 2025' and has been described as intensely private and low-profile.
- Known for: DeepSeek (founder & CEO, 2023-present); DeepSeek-R1 and DeepSeek-V3 models; High-Flyer Capital Management (co-founder, 2015)
- Education: Zhejiang University (浙江大学)

### Yu Wu (吴俣)
Head of LLM Alignment Team, DeepSeek

Yu Wu (吴俣) is Head of the Post-Training Team at DeepSeek. He earned his Ph.D. from Beihang University in 2019, supervised by Prof. Li Zhoujun and Dr. Zhou Ming (MSRA VP). He interned at MSRA (2013-2019), working on XiaoIce and Bing Encyclopedia. He leads post-training for DeepSeek LLM, DeepSeek-Math, and DeepSeek-Coder series and is a core contributor to DeepSeek-R1 (Nature 2025).
- Known for: Head of Post-Training, DeepSeek; DeepSeek-R1 core contributor; DeepSeek LLM/Math/Coder post-training lead
- Education: Beihang University (北京航空航天大学)

### Daya Guo (郭达雅)
Researcher, DeepSeek (reportedly departed for ByteDance Seed early 2026)

Daya Guo (郭达雅) is a Chinese AI researcher who was a core researcher at DeepSeek before reportedly joining ByteDance's Seed team in early 2026 as an Agent lead at L8 level. He was a lead researcher on DeepSeek's influential R1 model and co-authored numerous key DeepSeek papers including DeepSeek-Coder, DeepSeek-Math, DeepSeek-Prover, DeepSeek-V3, and DeepSeek-R1. Guo completed his Ph.D. at Sun Yat-sen University's School of Data and Computer Science, where he was jointly supervised by Professor Jian Yin of Sun Yat-sen and Dr. Ming Zhou of Microsoft Research Asia (MSRA). During his MSRA internship, he contributed to CodeBERT and GraphCodeBERT. He won three consecutive Tencent Advertising Algorithm Competition championships (2019-2021) and the Ant ATEC Technology Elite Competition. His Google Scholar citations exceed 3,800. His reported move to ByteDance Seed was one of the highest-profile AI talent moves in China in 2026.
- Known for: DeepSeek-R1 (lead researcher); DeepSeek-Coder series; DeepSeek-Math; DeepSeek-Prover; CodeBERT and GraphCodeBERT contributions
- Education: Sun Yat-sen University (中山大学)

### Zhihong Shao (邵智宏)
Research Scientist, DeepSeek

Zhihong Shao (邵智宏) is a Research Scientist at DeepSeek, recognized as one of MIT Technology Review's 35 Innovators Under 35. He received his Ph.D. in Computer Science from Tsinghua University, advised by Professor Minlie Huang. Shao's research focuses on LLM reasoning, tool augmentation, and mathematical problem solving. He is a key contributor to multiple landmark DeepSeek projects: DeepSeekMath (first author), DeepSeek-R1 (published in Nature), DeepSeek-Prover series, DeepSeek-Coder-V2, and DeepSeek-V3. He also led the ToRA project (tool-integrated reasoning agent), which was the first open-source model to exceed 50% accuracy on the competition-level MATH dataset. His work spans informal math pre-training, formal theorem proving, and inference-time optimization for LLMs.
- Known for: DeepSeekMath (first author); DeepSeek-R1 (Nature publication); DeepSeek-Prover series; DeepSeek-Coder-V2; ToRA: Tool-Integrated Reasoning Agent (ICLR 2024); Math-Shepherd (ACL 2024)
- Education: Beihang University (北京航空航天大学), Tsinghua University (清华大学)

### Zhibin Gou (苟志斌)
Core researcher, DeepSeek

Zhibin Gou (苟志斌) is a researcher at DeepSeek since June 2024. He holds an M.S. in AI from Tsinghua (advisor: Prof. Yujiu Yang, GPA 4.0) and B.S. in CS from BUPT (Rank 1%). He was a research intern at MSRA (NLC Group, 2023-2024). He is first author of CRITIC (ICLR 2024), co-first author of ToRA (ICLR 2024) and Rho-1 (NeurIPS 2024 Oral, Best Paper Runner-up), and a core contributor to DeepSeek-R1 (Nature 2025).
- Known for: DeepSeek researcher (Jun 2024); DeepSeek-R1 core contributor; CRITIC (ICLR 2024, 1st author); ToRA (ICLR 2024, co-1st); Rho-1 (NeurIPS 2024 Oral, Best Paper Runner-up)
- Education: Beijing University of Posts and Telecommunications (北京邮电大学), Tsinghua University (清华大学)

### Dejian Yang (杨德建)
Researcher, DeepSeek

Dejian Yang (杨德建) is a core researcher at DeepSeek. He is one of four consistent core contributors across DeepSeek's major papers (with Daya Guo, Qihao Zhu, Zhihong Shao), contributing to DeepSeek-R1 (Nature 2025) and DeepSeek-V3.
- Known for: Core researcher, DeepSeek; DeepSeek-R1 core contributor (Nature 2025); DeepSeek-V3 core contributor
- Education: Beihang University (北京航空航天大学)

### Junxiao Song (宋俊潇)
Principal Researcher, DeepSeek

Junxiao Song (宋俊潇) is a Principal Researcher at DeepSeek. He is a co-author of the landmark DeepSeek-R1 paper published in Nature, which demonstrated how reinforcement learning can incentivize reasoning capabilities in large language models. Song was previously a Ph.D. researcher at the Hong Kong University of Science and Technology (HKUST), in the Department of Electronic and Computer Engineering. His research at DeepSeek focuses on reinforcement learning for reasoning, mathematical reasoning in LLMs, and automated theorem proving. He has contributed to multiple DeepSeek model releases including DeepSeek-R1, DeepSeek-Prover, and DeepSeek-Prover-V2.
- Known for: DeepSeek-R1 (co-author, published in Nature); DeepSeek-Prover series; DeepSeek-Prover-V2
- Education: Zhejiang University (浙江大学), Hong Kong University of Science and Technology

### Damai Dai (代达劢)
Deep Learning Researcher, DeepSeek

Damai Dai (代达劢) is a Deep Learning Researcher at DeepSeek, where he serves as the infrastructure lead responsible for inference system engineering optimization and large-scale deployment. He earned his Ph.D. in 2024 from the School of Computer Science at Peking University, under the supervision of Professor Sui Zhifang at the MOE Key Lab of Computational Linguistics. Dai has published over 20 top conference papers with more than 28,000 Google Scholar citations. In 2023, he won the EMNLP Best Long Paper Award as the third core author for 'Label Words are Anchors' - the first time a mainland Chinese institution won this award. He is the first author of the DeepSeekMoE paper (ACL 2024), which introduced the Mixture-of-Experts architecture that underpins DeepSeek's cost-efficient model performance. He has been involved in DeepSeek's development from V1 through V3.
- Known for: DeepSeekMoE (first author, ACL 2024); DeepSeek infrastructure and inference optimization; EMNLP Best Long Paper Award (2023)
- Education: Peking University (北京大学)

### Qihao Zhu (朱启昊)
Researcher, DeepSeek

Qihao Zhu (朱启昊) is a researcher at DeepSeek focusing on code generation. He earned his Ph.D. from Peking University under Prof. Yingfei Xiong. He has published at ICSE, IJCAI, and AAAI on program generation and code repair.
- Known for: DeepSeek researcher (code generation); Ph.D. PKU (advisor: Prof. Yingfei Xiong); Publications at ICSE, IJCAI, AAAI
- Education: Peking University (北京大学)

### Tang Jie (唐杰)
Co-founder & controlling shareholder, Zhipu; WeBank Chair Professor, Tsinghua CS

Tang Jie (唐杰, born c. 1977) is the co-founder and controlling shareholder of Zhipu AI (智谱), a leading Chinese AI foundation model company that went public on the Hong Kong Stock Exchange in January 2026 as the world's first publicly listed LLM company. He is also a WeBank Chair Professor at Tsinghua University's Department of Computer Science. Tang launched Zhipu in 2019 with Liu Debing, who serves as chairman. The company was added to the U.S. government's Entity List in March 2025. Tang is a leading researcher in artificial general intelligence, data mining, social networks, and knowledge graphs. Zhipu's GLM model series serves over 12,000 enterprise clients and 45 million developers.
- Known for: Zhipu AI (co-founder & controlling shareholder, 2019-present); GLM model series; Zhipu IPO on HKEX (Jan 2026) - world's first public LLM company
- Education: Tsinghua University (清华大学)

### Li Juanzi (李娟子)
Co-founder, Zhipu; Professor, Tsinghua CS (KEG)

Li Juanzi (李娟子, also known as Li Juanzi) is a Professor at Tsinghua University's Department of Computer Science and Technology and a co-founder of Zhipu AI. She founded the Knowledge Engineering Group (KEG) laboratory at Tsinghua in 2001, which became the foundational research group for Chinese knowledge graph technology. Her students include Zhang Peng (now Zhipu CEO) and Tang Jie (now Zhipu co-founder and Tsinghua professor). Her research focuses on semantic web, semantic web services, news mining, and social network analysis. She earned her B.S. and M.S. from Shanxi University and her Ph.D. from Tsinghua in 2000. She has served as Deputy Director of the Institute of Software and Theory at Tsinghua and Deputy Director of the Chinese Information Processing Technology Committee of the China Computer Federation.
- Known for: Founded Tsinghua KEG Lab (2001); Co-founder of Zhipu AI; CN-DBpedia Chinese knowledge graph
- Education: Shanxi University (山西大学), Tsinghua University (清华大学)

### Zhang Peng (张鹏)
CEO, Zhipu

Zhang Peng (张鹏) is the CEO of Zhipu AI (智谱), the Chinese foundation model company that became one of the first publicly listed LLM companies when it IPO'd on the Hong Kong Stock Exchange in January 2026. He co-founded Zhipu in 2019 with Wang Shaolan and others, spinning it out of Tsinghua University's Knowledge Engineering Lab (KEG). Zhang joined Tsinghua in 1998, completing his undergraduate and master's degrees in Computer Science before earning his Ph.D. in 2018. Under his leadership, Zhipu built a full-stack model portfolio spanning language, multimodal systems, and AI agents, serving over 12,000 enterprise clients. The company raised HK$4.35 billion in its IPO at a valuation close to $9 billion; its current public-market valuation is tracked dynamically in the Atlas money view.
- Known for: Zhipu AI (CEO, 2019-present); GLM model series; Zhipu IPO - world's first public LLM company (Jan 2026)
- Education: Tsinghua University (清华大学)

### Yuxiao Dong (董昱潇)
Tsinghua KEG Associate Professor; senior collaborator on GLM models

Yuxiao Dong (董昱潇) is an Associate Professor at Tsinghua University's Knowledge Engineering Group (KEG). He earned his Ph.D. in Computer Science from the University of Notre Dame in 2017. Before joining Tsinghua, he was a researcher at Meta AI and Microsoft Research Redmond. His research focuses on data mining, graph machine learning, and large language models. He has contributed to GLM-130B, WebGLM, CodeGeeX, ChatGLM, and other models. He received the 2022 ACM SIGKDD Rising Star Award, the 2017 ACM SIGKDD Doctoral Dissertation Award Honorable Mention, and the ECML-PKDD 2023 Best Student Paper Award. His work on graph representation learning has been deployed at Facebook, Microsoft, Tencent, and Alibaba.
- Known for: Associate Professor, Tsinghua KEG; Senior collaborator on GLM models
- Education: University of Notre Dame

### Zhengxiao Du (杜正晓)
Tech Lead, Zhipu (since January 2023)

Zhengxiao Du (杜正晓) is a Tech Lead at Zhipu and a core contributor to the GLM model series. He is listed as a core contributor on the GLM-4.5 technical report alongside colleagues including Bin Chen, Cunxiang Wang, and Yushi Bai. He has been a tech lead at Zhipu since January 2023, contributing to the development of the company's large language models including the GLM-4 series.
- Known for: Tech Lead, Zhipu (since Jan 2023); GLM-4.5 core contributor
- Education: Tsinghua University (清华大学)

### Ming Ding (丁铭)
Now at ByteDance Seed (formerly Zhipu)

Ming Ding (丁铭) is a researcher at ByteDance Seed, having previously worked at Zhipu AI. At Zhipu, he was one of the founding team core members and served as the video/visual model lead. He was a core author of CogVLM (visual large language model) and deeply contributed to the development of CogView (text-to-image model). He was part of Zhipu's visual model research team before transitioning to ByteDance Seed.
- Known for: ByteDance Seed researcher (formerly Zhipu); Zhipu founding team core member; Video/visual model lead, Zhipu; Core author of CogVLM
- Education: Tsinghua University (清华大学)

### Yang Zhilin (杨植麟)
Founder & CEO, Moonshot AI

Yang Zhilin (杨植麟, born 1992) is the founder and CEO of Moonshot AI (月之暗面), the Beijing-based startup behind the Kimi large language model family. Founded in April 2023, Moonshot AI was valued around $3.3 billion in late 2024 and later reportedly raised $2 billion at a $20 billion valuation in May 2026, backed by investors including Alibaba, Tencent, and Sequoia China/HongShan. Yang earned his bachelor's degree from Tsinghua University (2015, top of his CS class) and completed his Ph.D. at Carnegie Mellon University in under four years, advised by Ruslan Salakhutdinov. He has published over 20 influential AI papers and co-authored XLNet (cited over 10,000 times) with Turing Award winners Yoshua Bengio and Yann LeCun. He previously worked at Google Brain and Meta AI Research. In July 2025, Moonshot AI released and open-sourced Kimi K2, China's first open-source trillion-parameter model.
- Known for: Moonshot AI / Kimi (founder & CEO, 2023-present); Kimi K2 (trillion-parameter open-source model, 2025); XLNet (co-author, 10,000+ citations); Transformer-XL (co-first author)
- Education: Tsinghua University (清华大学), Carnegie Mellon University

### Zhang Yutao (张宇韬)
Co-founder & CTO, Moonshot AI

Zhang Yutao (张宇韬) is the co-founder and CTO of Moonshot AI. In his early 30s, he oversees the company's technical architecture and model development. He holds a Ph.D. in Computer Science from Tsinghua University, where he was advised by Professor Tang Jie, the co-founder of Zhipu. He was a classmate of Yang Zhilin and Zhou Xinyu at Tsinghua's Computer Science Department (Class of 2011 undergraduate). Moonshot AI's chatbot Kimi has more than 50 million monthly active users across its app and web editions; the company's valuation moved from roughly $3.3 billion in late 2024 reports to a reported $20 billion financing mark in May 2026.
- Known for: Moonshot AI co-founder & CTO; Kimi chatbot technical architecture
- Education: Tsinghua University (清华大学)

### Zhou Xinyu (周昕宇)
Co-founder & head of algorithms, Moonshot AI

Zhou Xinyu (周昕宇) is a co-founder and head of algorithms at Moonshot AI, holding a 10% stake in the company. He was a classmate of Yang Zhilin and Zhang Yutao in Tsinghua University's Computer Science Department (Class of 2011 undergraduate). In his senior year, he interned at Megvii and joined full-time after graduation, focusing on algorithm mass production. He co-authored the ShuffleNet paper (accepted by CVPR) with Zhang Xiangyu of Megvii Research as co-first authors. ShuffleNet later influenced millisecond-level face unlock technologies including Apple's 3D Face Unlock. Zhou and Yang Zhilin were also bandmates in a rock band called Splay at Tsinghua.
- Known for: Moonshot AI co-founder (10% stake); ShuffleNet (co-first author, CVPR); Head of algorithms at Moonshot AI
- Education: Tsinghua University (清华大学)

### Wu Yuxin (吴育昕)
Co-founder & research lead, Moonshot AI

Wu Yuxin (吴育昕) is a co-founder and research lead at Moonshot AI (月之暗面). He graduated from both Tsinghua University and Carnegie Mellon University (like Yang Zhilin), with a research focus on computer vision detection and recognition. After graduation, he worked at Meta's Fundamental AI Research (FAIR) lab, where he co-proposed Group Normalization (GN) with Kaiming He and created Detectron2, one of Meta's most popular AI projects. His papers have received over 19,000 Google Scholar citations, and he was nominated for Best Paper at ECCV 2018. In 2018, his team IYSWIM was the only one to crack a facial recognition algorithm at the GeekPwn international security competition.
- Known for: Moonshot AI co-founder; Group Normalization (GN) co-author with Kaiming He; Detectron2 creator; 19,000+ Google Scholar citations
- Education: Tsinghua University (清华大学), Carnegie Mellon University

### Su Jianlin (苏剑林)
Researcher, Moonshot AI

Su Jianlin (苏剑林) is a researcher at Moonshot AI, known as the 'God of Math' (苏神) in the Chinese AI community. He is the originator of RoPE (Rotary Position Embedding), a widely adopted positional encoding technique used in many large language models including LLaMA and the Qwen series. At Moonshot AI, he contributed to the Kimi Linear architecture and the Attention Residuals paper that impressed Elon Musk. He has published 46 papers with over 7,700 citations and an h-index of 23. He maintains the popular science blog kexue.fm. At Moonshot, he co-authored a paper on Attention Residuals with a 17-year-old high school student (first author) and Zhang Yu (Kimi Linear first author), which improved training efficiency by 25%.
- Known for: RoPE (Rotary Position Embedding) - originator; Attention Residuals (co-author, 2026); Kimi Linear architecture contributor; 46 papers, 7,700+ citations
- Education: Sun Yat-sen University (中山大学)

### Yan Junjie (闫俊杰)
Founder, Chairman, CEO and CTO, MiniMax

Yan Junjie (闫俊杰, born 1989) is the Founder, Chairman, CEO, and CTO of MiniMax (稀宇科技), one of China's leading AI companies. Born in a small county in Henan province, he taught himself advanced calculus in high school. He earned his B.S. from Southeast University (2006-2010) and his Ph.D. from the Chinese Academy of Sciences Institute of Automation (2015), followed by postdoctoral work at Tsinghua University. He joined SenseTime in 2015, rising to Vice President and CTO of the Smart Cities business group, leading 700+ people. In late 2021, he left SenseTime to found MiniMax, which IPO'd on the Hong Kong Stock Exchange on January 9, 2026, becoming one of the first listed large-model companies. MiniMax's public-market valuation has moved quickly since listing; current market cap is tracked dynamically in the Atlas money view. MiniMax serves 2.36 billion users across 200+ countries with products including Hailuo AI, Talkie, and MiniMax Agent.
- Known for: MiniMax (founder, chairman, CEO & CTO, 2021-present); MiniMax IPO on HKEX (Jan 2026); SenseTime VP & CTO Smart Cities (2015-2021)
- Education: Southeast University (东南大学), Institute of Automation, CAS (中科院自动化所), Tsinghua University (清华大学)

### Zhou Yucong (周彧聪)
Co-founder; Executive Director and Visual Model Research Leader, MiniMax

Zhou Yucong (周彧聪, born ~1993, aged 32 in 2025) is a co-founder, Executive Director, and the Visual Model Research Leader at MiniMax. He joined MiniMax in March 2022 as a visual model researcher and engineer. Previously, he worked at SenseTime (April 2018 - July 2019) and Huawei (August 2019 - March 2022), focusing on algorithms. He holds B.S. and M.S. degrees from Beihang University. He was appointed as Executive Director in June 2025. He leads MiniMax's visual model R&D including the Hailuo AI video generation product.
- Known for: Co-founder and Executive Director, MiniMax; Visual Model Research Leader, MiniMax; Hailuo AI video generation
- Education: Beihang University (北京航空航天大学)

### Yang Bin (杨斌)
(former) Co-founder, MiniMax — departed 2024

Yang Bin (杨斌) was a co-founder of MiniMax and served as the head of model training. He was a classmate of Yan Junjie at the University of Chinese Academy of Sciences (UCAS). He earned his master's degree from the Institute of Automation, Chinese Academy of Sciences in 2014, where he worked on deep learning projects. He then pursued a Ph.D. at the University of Toronto (2016-2022) under Professor Raquel Urtasun, a senior research scientist at Uber ATG, where he helped build Uber's autonomous driving AI R&D center from scratch. During his Ph.D., he co-founded Waabi, an autonomous trucking company, with his advisor. After returning to China, he co-founded MiniMax with Yan Junjie and led the technical team to develop MiniMax's foundational models across text-to-visual, text-to-audio, and text-to-text modalities. He was the second founding core member to leave MiniMax after Song Yachen departed in late 2022.
- Known for: Co-founder of MiniMax (departed Jul 2024); Co-founded Waabi (autonomous trucking); Built Uber's autonomous driving AI R&D center
- Education: University of Science and Technology of China (中国科学技术大学), Institute of Automation, CAS (中科院自动化所), University of Toronto

### Yun Yeyi (贠烨祎)
Executive Director and COO, MiniMax

Yun Yeyi (贠烨祎, born 1994) is the Executive Director and Chief Operating Officer (COO) of MiniMax. She holds approximately 3.93% of MiniMax's shares, with a stake valued at approximately HK$4.8 billion (about $600M USD) at IPO. She earned her B.S. in Electronic Engineering from Johns Hopkins University in 2017, with minors in Economics and Mathematics. She joined SenseTime in 2017 as a manager of fundraising and strategic investment, was promoted to CEO Executive Assistant and Strategy Director in 2018, and then served as Director of Innovative Business from 2021-2022 before co-founding MiniMax with Yan Junjie. She is described as 'capable, commanding, and execution-strong' with maturity beyond her age.
- Known for: Executive Director and COO, MiniMax; Co-founded MiniMax (2021)
- Education: Johns Hopkins University

### Zhao Pengyu (赵鹏宇)
Executive Director and LLM Research Leader, MiniMax (June 2025)

Zhao Pengyu (赵鹏宇, born ~1995, aged 29 in 2025) is the Executive Director and Large Language Model Research Leader at MiniMax. He joined MiniMax in August 2023 as an NLP researcher and engineer. He holds a B.S. (2017) and M.S. (2020) in Computer Science from Peking University. Before joining MiniMax, he worked at Beijing Hulu Technology as a research software development engineer focusing on recommendation algorithms (August 2020 - July 2023). He was appointed Executive Director in June 2025 and leads MiniMax's LLM research and engineering efforts. He is also a co-author of MiniMax's MiniMax-01 technical report.
- Known for: Executive Director and LLM Research Leader, MiniMax; Co-author MiniMax-01 technical report
- Education: Peking University (北京大学)

### Yongyi Hu (胡咏毅)
(former) Core algorithm engineer, MiniMax — now co-founder, Phronesis AI

Yongyi Hu (胡咏毅, also known as Pascal Hu) is the Chief Scientist and Co-founder of Phronesis AI, where he built EigenFlux.ai - an open-source agent-to-agent communication network with 3,200+ connected agents. He previously served as a Large Model Algorithm Lead at MiniMax (2023-2025), where he led model data, Talkie/Xingye product algorithms, and code pre-training across the full stack. Before MiniMax, he was a Research Intern at Meta AI (FAIR Paris), where he worked on automated theorem proving using LLM + reinforcement learning in the Lean environment and contributed to early-stage data collection for the LLaMA project. He holds dual degrees from Shanghai Jiao Tong University and Ecole Polytechnique (X) in France.
- Known for: Co-founder & Chief Scientist, Phronesis AI; EigenFlux.ai - agent communication network; MiniMax Large Model Algorithm Lead (2023-2025); Contributed to LLaMA at Meta FAIR Paris
- Education: Shanghai Jiao Tong University (上海交通大学), École Polytechnique

### Kai-Fu Lee (李开复)
Founder & CEO, 01.AI; Founder/Chairman, Sinovation Ventures

Kai-Fu Lee (李开复) is a globally renowned AI expert, venture capitalist, and author. He is the Founder and CEO of 01.AI, a Chinese generative AI company developing large language models, and the Chairman and CEO of Sinovation Ventures, a leading venture capital firm managing approximately $3 billion. Previously, Lee served as President of Google China (2005-2009) and held executive positions at Microsoft, SGI, and Apple. At Microsoft, he founded Microsoft Research China (later MSRA), recognized by MIT Technology Review as one of the world's top research institutions. He earned his B.S. from Columbia University and his Ph.D. in Computer Science from Carnegie Mellon University, where he developed one of the first speaker-independent continuous speech recognition systems. He is the author of bestselling books 'AI Superpowers' and 'AI 2041', and was named to TIME's list of 25 most influential AI leaders.
- Known for: 01.AI (founder & CEO, 2023-present); Sinovation Ventures (chairman & CEO, 2009-present); Microsoft Research China (founder); Google China (President, 2005-2009); Author of 'AI Superpowers' and 'AI 2041'
- Education: Columbia University, Carnegie Mellon University

### Wenhao Huang (黄文灏)
ByteDance large-model team; former 01.AI technical co-founder / VP Algorithms

Wenhao Huang (黄文灏) joined ByteDance's large-model team in August 2024, reportedly reporting to Zhu Wenjia. He earned his Ph.D. from Peking University CS, worked at MSRA on agent research, then led BAAI Health Computing Research Center. He was technical co-founder and VP of Algorithms at 01.AI before leaving for ByteDance.
- Known for: ByteDance large-model team (joined Aug 2024); Former technical co-founder / VP Algorithms, 01.AI; MSRA agent research; BAAI Health Computing Research Center
- Education: Peking University (北京大学)

### Guoyin Wang (王国胤)
(former) Co-founder & Head of Post-Training, 01.AI — now leading Alibaba Tongyi Qwen Pilot group

Guoyin Wang (王国胤) is a former Co-founder and Head of Post-Training at 01.AI. Following the March 2026 restructuring of Alibaba's Tongyi Lab, he now leads the Qwen Pilot group within Alibaba. He was part of the team that departed 01.AI along with the broader talent movement to Alibaba. His work focuses on post-training and model alignment for large language models.
- Known for: Former Co-founder & Head of Post-Training, 01.AI; Now leads Qwen Pilot group at Alibaba

### Zonghong Dai (Ethan Dai) (戴宗宏)
(former) Co-founder, VP of Tech, Head of AI Infra, 01.AI — founded Jidian Qiyuan March 2025


- Known for: Former Co-founder, VP of Tech, Head of AI Infra at 01.AI; Founded Jidian Qiyuan (March 2025)

### Heng Ji (季姮)
Full Professor, UIUC; Amazon Scholar; Founding Director of Amazon-Illinois AICE Center

Heng Ji (季姮) is a Full Professor at the University of Illinois Urbana-Champaign (UIUC) Siebel School of Computing and Data Science, with affiliated appointments in Electrical and Computer Engineering and the Coordinated Science Laboratory. She is an Amazon Scholar and the Founding Director of the Amazon-Illinois Center on AI for Interactive Conversational Experiences (AICE). She earned her B.A. and M.A. in Computational Linguistics from Tsinghua University and her M.S. and Ph.D. in Computer Science from New York University. Her research focuses on Natural Language Processing, particularly multimedia multilingual information extraction, knowledge-enhanced large language models, and vision-language models. She was elected as an ACL Fellow in 2025. Her awards include NSF CAREER (2009), IEEE 'AI's 10 to Watch' (2013), two NAACL 2024 Outstanding Paper Awards, and the 2023 WLA Young Scientist award. She has testified to the U.S. House as an AI expert and leads multiple DARPA and ARL projects.
- Known for: Amazon-Illinois AICE Center (Founding Director); ACL Fellow (2025); NSF CAREER Award (2009); IEEE 'AI's 10 to Watch' (2013); Multiple DARPA project leadership
- Education: Tsinghua University (清华大学), New York University

### Bei Chen (陈蓓)
Researcher, MSRA

Bei Chen (陈蓓) is a Senior Researcher at Microsoft Research Asia (MSRA). She joined Microsoft in 2017 after receiving her Ph.D. from the TSAIL Group in the Department of Computer Science and Technology at Tsinghua University, where she was jointly supervised by Prof. Bo Zhang and Prof. Jun Zhu. She received her B.E. degree from the Department of Computer Science and Technology at Harbin Institute of Technology in 2012. Her research interests include machine learning and its applications in natural language processing and data mining, especially latent feature models, probabilistic graphical models, Bayesian nonparametrics, reinforcement learning, and deep learning.
- Known for: Senior Researcher at MSRA
- Education: Harbin Institute of Technology (哈尔滨工业大学), Tsinghua University (清华大学)

### Wang Xiaochuan (王小川)
Founder & CEO, Baichuan

Wang Xiaochuan (王小川, born 1978) is the founder and CEO of Baichuan AI (百川智能), a Chinese large language model startup. He is a well-known Chinese tech entrepreneur and computer scientist. Prior to founding Baichuan in 2023, Wang served as CEO of Sogou (搜狗), one of China's leading search engines, which was acquired by Tencent. He holds multiple degrees from Tsinghua University, including a B.S., M.S., and Ph.D. in Computer Science, as well as an EMBA. Wang is widely regarded as a 'tech geek' and is one of the most prominent figures in China's AI industry.
- Known for: Baichuan AI (founder & CEO, 2023-present); Sogou (former CEO, acquired by Tencent)
- Education: Tsinghua University (清华大学)

### Ru Liyun (茹立云)
Co-founder, Baichuan

Ru Liyun (茹立云) is a co-founder of Baichuan Intelligence. He was previously the COO of Sogou, where he worked alongside CEO Wang Xiaochuan. After Sogou was acquired by Tencent, Ru left and co-founded Baichuan with Wang Xiaochuan in April 2023. In 2018, he founded an AI education company called Putao Zhixue (葡萄智学) before returning to co-found Baichuan.
- Known for: Co-founder, Baichuan Intelligence (2023-present); Former COO of Sogou
- Education: Tsinghua University (清华大学)

### Hong Tao (洪涛)
(former) Co-founder, Baichuan — departed July 2024

Hong Tao (洪涛) is a former co-founder of Baichuan Intelligence who departed in July 2024. He was previously the Chief Marketing Officer (CMO) at Sogou, where he led Sogou to become the 6th largest internet advertising revenue company in China. He joined Sogou in 2005 and built the promotion systems for Sogou Input Method, Browser, and Search. He was one of the 'Four Heavenly Kings' of Sogou alongside CEO Wang Xiaochuan, COO Ru Liyun, and CTO Yang Hongtao. He joined Baichuan in July 2023 as co-founder and head of commercialization, quickly building the commercial team. His departure was officially announced in December 2024 as being due to 'personal reasons.' He also previously worked at Konka and Yahoo.
- Known for: Former co-founder and CMO, Baichuan Intelligence; Former CMO of Sogou

### Chen Weipeng (陈未鹏)
(former) Co-founder & head of foundation-model team, Baichuan — founded Yongyue Intelligence June 2025

Chen Weipeng (陈未鹏) is a former co-founder and head of the foundation-model team at Baichuan Intelligence. He was a founding team member at Baichuan and previously served as General Manager of Sogou Search R&D. He left Baichuan in March 2025 (along with co-founder Jiao Ke) and founded Yongyue Intelligence (永悦智能) in June 2025, focusing on AI coding. He graduated from Harbin Institute of Technology.
- Known for: Former co-founder & head of foundation-model team, Baichuan; Founded Yongyue Intelligence (Jun 2025); Former Sogou Search R&D GM
- Education: Harbin Institute of Technology (哈尔滨工业大学)

### Wang Bingning (王炳宁)
AI algorithm research lead, Tencent WeChat Search; former Head of Pre-training, Baichuan

Wang Bingning (王炳宁) is an AI algorithm research lead for Tencent WeChat Search and a former Head of Pre-training at Baichuan AI. He earned his Ph.D. from the Institute of Automation, Chinese Academy of Sciences, where his dissertation won an outstanding dissertation award from the Chinese Information Processing Society of China. Before Baichuan, he worked on search, recommendation, computational advertising, question answering, and large language models at Tencent and Sogou. His Baichuan-era work includes leadership on Baichuan pretraining and authorship on Baichuan-M1 medical LLM.
- Known for: AI algorithm research lead, Tencent WeChat Search; Former Head of Pre-training, Baichuan AI; Baichuan-M1 medical LLM (lead author); Large Chinese QA and retrieval datasets; Outstanding doctoral dissertation award, Chinese Information Processing Society of China
- Education: Institute of Automation, CAS (中科院自动化所)

### Jiang Daxin (姜大昕)
Co-founder & CEO, StepFun

Jiang Daxin (姜大昕) is the co-founder and CEO of StepFun (阶跃星辰), a Shanghai-based AI company focused on multimodal foundation models. He earned his Ph.D. in Computer Science from the State University of New York at Buffalo, and served as an Assistant Professor at Nanyang Technological University in Singapore before joining Microsoft Research Asia (MSRA) as a Principal Researcher in 2007. He later became a Corporate Vice President at Microsoft and served as Deputy Managing Director and Chief Scientist at Microsoft's Search Technology Center Asia (STCA). He founded StepFun in 2023 and was named an IEEE Fellow in 2024. He was also selected as one of the '2025 AI Figures of the Year' by China AI Festival.
- Known for: StepFun (co-founder & CEO, 2023-present); IEEE Fellow (2024); Microsoft VP and STCA Chief Scientist
- Education: University at Buffalo (SUNY)

### Jiao Binxing (焦斌星)
Co-founder; head of search systems, StepFun

Jiao Binxing (焦斌星) is a co-founder of StepFun and the company's data lead. He graduated from the University of Science and Technology of China (USTC) and was selected for the USTC-MSRA joint PhD program. After completing his Ph.D., he joined Microsoft as the lead of the Bing core search team, responsible for using data mining and NLP algorithms to optimize indexing and search quality. He also developed automatic high-quality site mining algorithms used for site ranking and indexing. In 2023, he co-founded StepFun with Jiang Daxin, serving as the data lead providing critical data support for the company's large model development.
- Known for: Co-founder & Data Lead, StepFun (2023); Former Microsoft Bing core search team lead; USTC-MSRA joint PhD program
- Education: University of Science and Technology of China (中国科学技术大学)

### Zhu Yibo (朱亦博)
Co-founder & CTO, StepFun

Zhu Yibo (朱亦博) is the co-founder and CTO of StepFun (阶跃星辰). He obtained his Ph.D. from UC Santa Barbara (UCSB), co-advised by Prof. Ben Y. Zhao and Prof. Heather Zheng, and received his bachelor's degree from Tsinghua University. He is a recipient of the Microsoft Research Ph.D. Fellowship (2015). His career spans Microsoft Research (where he pioneered RoCE technology for high-speed data transfer), ByteDance (where he built AI infrastructure from scratch, managing clusters of 10,000+ GPUs), and Google Cloud (where he led GPU product technology and supported Anthropic). At StepFun, he leads the AI Infrastructure team, which is treated as equally important as the algorithm team - a unique 'dual-engine' architecture among Chinese AI startups. StepFun's B+ round raised over 5 billion RMB (~$690M+) in January 2026, setting a record for the largest single financing round in China's AI sector in the past 12 months.
- Known for: StepFun co-founder & CTO; RoCE technology pioneer at Microsoft Research; Built ByteDance AI Infra from scratch (10,000+ GPU clusters); Microsoft Research Ph.D. Fellowship 2015
- Education: Tsinghua University (清华大学), UC Santa Barbara

### Yin Qi (印奇)
Chairman, StepFun (January 2026); Chairman, Megvii; Chairman, Qianli Tech (Geely)

Yin Qi (印奇, born 1988) is the Chairman of StepFun (阶跃星辰) since January 2026, Chairman of Megvii (旷视科技), and Chairman of Qianli Technology (千里科技), Geely's autonomous driving subsidiary. He is a prominent figure in China's AI industry who co-founded Megvii in 2011, one of the 'AI Four Dragons' (四小龙) specializing in computer vision and Face++. After Megvii's IPO attempts (Hong Kong 2019, STAR Market 2021-2024) were unsuccessful and the company withdrew its application in December 2024, Yin pivoted to lead StepFun. He holds degrees from Tsinghua University (Yao Class/姚班) and Columbia University (master's). In January 2026, StepFun announced over 5 billion yuan (~$690M+) B+ round funding with Yin as Chairman, overseeing strategy, capital, and commercialization while CEO Jiang Daxin leads R&D.
- Known for: Megvii (co-founder, 2011-present; chairman); StepFun (chairman since Jan 2026); Qianli Technology (chairman since Nov 2024); Face++ computer vision platform
- Education: Tsinghua University (清华大学), Columbia University

### Gang Yu (余刚)
Principal Researcher AI Lab, StepFun (since 2024)

Gang Yu (余刚) is a Principal Research Scientist at StepFun (阶跃星辰). He earned his Ph.D. from Nanyang Technological University (NTU) in 2014, supervised by Prof. Junsong Yuan. Before joining StepFun, he was a Research Director at Tencent for four years and spent five years at Megvii (Face++). His research focuses on computer vision and generative AI, including object detection, segmentation, human keypoint detection, human action recognition, and 3D reconstruction. He has served as an Area Chair for CVPR 2024/2025/2026, ICCV 2025, and NeurIPS 2025, and as Innovation Program (Industry) Chair for ICME 2024. His team has published extensively at CVPR, ICCV, NeurIPS, ICLR, and ECCV.
- Known for: Principal Research Scientist, StepFun (since 2024); Former Research Director, Tencent (4 years); Former Megvii/Face++ (5 years); CVPR/ICCV/NeurIPS Area Chair

### Zhou Jingren (周靖人)
CTO Alibaba Cloud Intelligence; head of Tongyi Lab; Deputy Head of DAMO; Chief AI Architect

Zhou Jingren (周靖人) is a senior technology executive at Alibaba Group. He joined Alibaba as Chief Scientist at its cloud unit and later served as CTO of Alibaba Cloud Intelligence, overseeing the company's cloud and AI technology strategy. In April 2026, Alibaba underwent a major reorganization: Zhou stepped down as CTO (replaced by Li Feifei) but was elevated to lead the newly upgraded Tongyi Large Model Business Division (通义大模型事业部), effectively becoming Alibaba's top executive for model development. He also serves as Chief AI Architect on Alibaba's Technology Committee. Zhou holds a Ph.D. in Computer Science from Columbia University (2004) and previously served as a Partner-level Researcher at Microsoft. His career spans across cloud computing, algorithms, and business at Alibaba, making him a key cross-functional technology leader.
- Known for: Alibaba Cloud CTO; Head of Tongyi Lab / Tongyi Large Model Business Division; Alibaba Partnership member (since Oct 2025); Chief AI Architect, Alibaba Technology Committee
- Education: University of Science and Technology of China (中国科学技术大学), Columbia University

### Lin Junyang (林俊旸)
(former) Tech Lead, Qwen — resigned March 3, 2026

Lin Junyang (林俊旸) was the Tech Lead of Alibaba's Qwen team before resigning on March 3, 2026. His departure triggered a significant restructuring at Alibaba's Tongyi Lab. Following his resignation, Zhou Hao (formerly of Google DeepMind) was appointed as the new Qwen tech lead, and Zhou Jingren was elevated to head the newly upgraded Tongyi Large Model Business Division. Lin's departure was part of a broader wave of Qwen team exits in early 2026, which included his subordinate Bowen Yu (who joined ByteDance Seed), Binyuan Hui (who joined Meta), and others.
- Known for: Former Tech Lead, Qwen team (resigned Mar 3, 2026)
- Education: Peking University (北京大学)

### Fei Huang (黄飞)
Chief Scientist & Senior Director, Language Technologies Lab, DAMO Academy

Fei Huang (黄飞, also known as 黄非) is the Chief Scientist and Senior Director of the Language Technologies Lab at Alibaba DAMO Academy. He leads NLP foundational technology, dialogue systems, and machine translation R&D teams, which developed the AliNLP platform serving hundreds of internal and external clients. Prior to Alibaba, he led machine translation development at Facebook AML, helping reduce language barriers for billions of users. He was also a senior researcher at IBM Watson after graduating from Carnegie Mellon University with a Ph.D. focusing on language technologies. He has published 30+ papers and holds 20+ patents on machine translation and multilingual NLP. He has served as Area Chair at ACL-IJCNLP 2015 and NLPCC 2018.
- Known for: AliNLP platform (Alibaba); Facebook AML machine translation systems; IBM Watson senior researcher

### Hao Zhou (周浩)
New Qwen tech lead (March 2026), replacing Lin Junyang

Hao Zhou (周浩) is the new technical lead of the Qwen team at Alibaba, appointed in March 2026 to replace Lin Junyang. He previously served as a Senior Research Scientist at Google DeepMind, where he participated in the key development of Gemini 3.0. His appointment was announced alongside Alibaba's response to Lin Junyang's departure, and he was tasked with leading the post-training team for Qwen.
- Known for: Qwen technical lead (since Mar 2026); Gemini 3.0 development at Google DeepMind (former)

### Bowen Yu (郁博文)
Now at ByteDance Seed (joined March 12, 2026 as post-training lead, Visual Model & Multimodal Interaction Team)

Bowen Yu (郁博文) is the post-training lead for the Visual Model & Multimodal Interaction Team at ByteDance Seed, a role he assumed on March 12, 2026. Previously, he was an algorithm scientist at Alibaba Tongyi Laboratory and the post-training lead of the Qwen team, where he led development of the Qwen series Chat models that repeatedly ranked in the global top 10 on LMSYS Chatbot Arena. He earned his Ph.D. from the Institute of Information Engineering, Chinese Academy of Sciences in 2022, where he received the Chinese Academy of Sciences President's Award. He joined Alibaba DAMO Academy in 2022 through the 'Alibaba Star' campus recruitment program. Yu has published over 50 papers at top-tier conferences including ICML, ICLR, and ACL, with 2,000+ citations. He served as Area Chair for ACL and EMNLP. He departed Alibaba on March 4, 2026 along with his supervisor Lin Junyang, and was succeeded by Zhou Hao (formerly of Google DeepMind).
- Known for: ByteDance Seed post-training lead (since Mar 2026); Qwen series Chat models post-training lead; 50+ papers at ICML, ICLR, ACL (2000+ citations)
- Education: Central South University (中南大学), Institute of Information Engineering, CAS (中科院信息工程研究所)

### Shuai Bai (白帅)
Lead author Qwen3-VL; head of Qwen vision-language line

Shuai Bai (白帅) is a Senior Algorithm Expert at Alibaba's Qwen team, where he leads the Qwen vision-language model line. He is the lead author of the Qwen3-VL Technical Report. He graduated from Beijing University of Posts and Telecommunications (BUPT). His research areas include visual representation, multimodal learning, and visual unified understanding and generation. He has published numerous papers at top conferences with over 9,500 citations. He is responsible for developing the Qwen-VL series of models, which have become among the most widely used open-source vision-language models globally.
- Known for: Qwen3-VL Technical Report (lead author); Qwen-VL series models; 9,500+ citations

### An Yang
Lead author Qwen3 Technical Report; pre-training lead

An Yang is a core contributor and the lead author of the Qwen3 Technical Report. He is affiliated with the Qwen Team at Alibaba and Peking University, where his research focuses on natural language processing. According to Google Scholar, his work has accumulated approximately 29,000 citations. An Yang is listed as a core contributor alongside Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, and others on the Qwen3 Technical Report published in May 2025. His work centers on pre-training leadership for the Qwen series of large language models at Alibaba's Tongyi Lab.
- Known for: Qwen3 Technical Report (lead author, 2025); Qwen series pre-training leadership

### Wu Yonghui (吴永辉)
Head of Seed foundational research, ByteDance (since February 2025)

Wu Yonghui (吴永辉) is the Head of foundational research at ByteDance Seed, a position he assumed in February 2025. He previously served as Vice President of Research at Google DeepMind, where he was involved in the Gemini large model project and had a 17-year career at Google. Wu is an alumnus of Nanjing University. At Google, he led the neural machine translation system GNMT and participated in Gemini's research. Since joining ByteDance, he reports directly to CEO Liang Rubo and has restructured Seed's research organization, focusing on improving base model capabilities and building a world-class AI research team. Under his leadership, Seed has maintained around 1,500 staff and developed Doubao 2.0, a trillion-parameter multimodal model.
- Known for: ByteDance Seed Head of foundational research (since Feb 2025); Google DeepMind VP of Research (former); GNMT (Google Neural Machine Translation) leader
- Education: Nanjing University (南京大学), UC Riverside

### Zhu Wenjia (朱文佳)
Leads Doubao large-model applications, ByteDance

Zhu Wenjia (朱文佳) leads Doubao large-model applications at ByteDance. He joined ByteDance in 2015 from Baidu where he served as a search department architect. He was recognized as one of the 'Top 3 algorithm talents' at Toutiao. In June 2019, he was promoted to head of Toutiao, and in February 2021, he became TikTok's product and technology lead, reporting to TikTok CEO Zhou Shouzi. In 2023, he led the formation of ByteDance's large model team, focusing on language and image modalities. After Wu Yonghui joined in February 2025 as head of Seed foundational research, Zhu's reporting line was adjusted - he now reports to Wu Yonghui instead of directly to CEO Liang Rubo. He is based in Singapore.
- Known for: Leads Doubao large-model applications at ByteDance; Former head of Toutiao (2019); Former TikTok product and technology lead (2021); Founded ByteDance's large model team (2023)

### Wang Mingxuan (王明轩)
Senior Researcher, ByteDance Seed; Co-Director SIA Lab (joint w/ Tsinghua AIR)

Wang Mingxuan (王明轩) is a Senior Researcher at ByteDance Seed and Co-Director of SIA Lab, a joint research laboratory between ByteDance Seed and Tsinghua University's AIR (Institute for AI Industry Research). He is one of the key researchers identified as part of the '8 key people' supporting ByteDance's large models. His work focuses on large model training and reinforcement learning. He co-leads research efforts at SIA Lab with Zhou Hao from Tsinghua AIR, working on projects like MemAgent - a reinforcement learning-based memory agent for long-text processing.
- Known for: Senior Researcher, ByteDance Seed; Co-Director SIA Lab (ByteDance Seed + Tsinghua AIR); MemAgent long-text memory agent

### Zhou Chang (周畅)
Multimodal Interaction & World-Models lead, ByteDance Seed (joined September 2024)

Zhou Chang (周畅) leads multimodal interaction and world-models at ByteDance Seed, having joined in September/October 2024. Previously, he spent seven years at Alibaba (2017-2024) where he served as the technical lead for the Qwen large model series. Under his leadership, Qwen grew from its April 2023 debut to become the world's most downloaded open-source model family, with over 20 million derivative models and 1 billion+ downloads on Hugging Face. In Qwen2 (June 2024), the model surpassed Llama 3-70B and topped the Hugging Face leaderboard within two hours. At ByteDance, he was assigned to the 'Multimodal Interaction & World-Models' team within Seed, and later took over Seedream (image generation) and Seedance (video generation) after Yang Jianchao took leave and Jiashi Feng departed. He holds a Bachelor's from Fudan University and a Ph.D. from Peking University. His reported compensation at ByteDance was an eight-figure annual package at level 4-2/5-1.
- Known for: ByteDance Seed multimodal lead (since Sep 2024); Qwen series technical lead at Alibaba (2017-2024); Seedance 2.0 video generation model (2026)
- Education: Fudan University (复旦大学), Peking University (北京大学)

### Yang Jianchao (杨建超)
(former) Multimodal Foundation Model Head, ByteDance Seed — took leave 2025

Yang Jianchao (杨建超) was the head of visual multimodal generation at ByteDance Seed, responsible for the Doubao large model's visual multimodal direction. He announced a temporary leave in July 2025, with his position being taken over by Zhou Chang (formerly of Alibaba's Qwen team). Prior to ByteDance, Yang worked as a visual algorithm researcher at Adobe and Snapchat. He joined ByteDance in 2018 and took full responsibility for visual model development since 2023. His leave was attributed to difficulties balancing work rhythms between North America and China, long-term pressure, and family reasons.
- Known for: Former head of visual multimodal generation, ByteDance Seed; Led Doubao visual model development since 2023

### Jiashi Feng (冯佳时)
(former) Visual Foundation Model lead, ByteDance Seed — departed 2025

Jiashi Feng (冯佳时) was the head of visual foundation model research at ByteDance Seed, where he led the visual basic research team for the Doubao large model. He departed ByteDance in August 2025. He received his B.S. from the University of Science and Technology of China (USTC) in 2007, his M.S. from the Institute of Automation, Chinese Academy of Sciences, and his Ph.D. from the National University of Singapore (NUS) in 2014. From 2014-2015, he was a postdoc at UC Berkeley, and from 2016-2021, he served as an Assistant Professor at NUS, where he succeeded his mentor Yan Shuicheng as head of the Machine Learning and Vision Lab. He joined ByteDance in 2019 (some sources say 2021) and became one of the 8 key figures in the company's Top Seed talent program. He has published over 400 papers, received MIT Technology Review 'Innovators Under 35' Asia award, ACM MM Best Student Paper Award, ICCV TASK-CV Best Paper Award, and served as Area Chair for CVPR, ICML, ICLR, and NeurIPS.
- Known for: Led Doubao visual foundation model research at ByteDance Seed; 400+ published papers; MIT TR35 Asia; MagicAnimate video model
- Education: University of Science and Technology of China (中国科学技术大学), National University of Singapore

### Qiao Mu (乔木)
(former) Doubao Seed team head — dismissed June 2025 over ethics probe

Qiao Mu (乔木) was the head of ByteDance's Doubao large language model team in the Seed department. He was dismissed by ByteDance in June 2025 following an ethics probe. An internal investigation found that he had an intimate relationship with an HRBP supporting his team, which constituted a conflict of interest under company policy. Both parties failed to declare the conflict and made false statements during the investigation. Qiao had been at ByteDance since 2014, having previously worked at Baidu. He held a B.S. and M.S. from Xi'an Jiaotong University. At ByteDance, he successively served as a recommendation algorithm engineer, head of Toutiao recommendations, and head of search technology before leading the Doubao LLM team. He reported to Wu Yonghui after the February 2025 reorganization.
- Known for: Former head of Doubao LLM team, ByteDance Seed (dismissed Jun 2025)

### Yao Shunyu (姚顺雨)
Chief AI Scientist, Tencent (December 2025)

Yao Shunyu (姚顺雨, born 1998, also known as Vinces Yao) is the Chief AI Scientist at Tencent, appointed in December 2025. He is a prodigious AI researcher who completed his undergraduate studies at Tsinghua University's prestigious Yao Class (姚班) and earned his Ph.D. in Computer Science from Princeton University in 2024, specializing in natural language processing and reinforcement learning. At Princeton, he proposed the Tree of Thoughts (ToT) framework and the ReAct method, which became foundational paradigms for AI agent development. He joined OpenAI in August 2024, where he was a core contributor to Operator and Deep Research. In May 2025, he was named to MIT Technology Review's '35 Innovators Under 35' China list. At Tencent, he serves as Chief AI Scientist in the CEO/President Office, reporting directly to President Martin Lau, while also heading the AI Infra Department and Large Language Model Department. In April 2026, he delivered Tencent's first major model release under his leadership: Hy3 preview (295B parameters), which was open-sourced.
- Known for: ReAct framework (co-author); Tree of Thoughts (ToT) framework; Tencent Hy3 preview model (Apr 2026); OpenAI Operator and Deep Research (core contributor)
- Education: Tsinghua University (清华大学), Princeton University

### Lu Shan (卢山)
President, Tencent TEG (Technology and Engineering Group)

Lu Shan (卢山) is the Senior Executive Vice President and President of Tencent's Technology and Engineering Group (TEG), the organizational base for Tencent's foundational AI functions including AI Lab and AI infrastructure. He joined Tencent in 2000 and has served in various roles including General Manager of the IM Product Division, Vice President of the Platform R&D System, and Senior Vice President of the Operations Platform System. Since May 2012, he has led TEG. He is a member of Tencent's executive council (总办), the company's highest decision-making body. Prior to joining Tencent, he worked at Shenzhen Liming Network Systems Limited. TEG is responsible for AI R&D, cloud infrastructure, data platforms, and foundational technology across Tencent's business groups.
- Known for: President of Tencent TEG (since 2012); Member of Tencent executive council (总办)

### Liu Yuhong (刘煜宏)
VP Tencent Cloud; Head of AI Data Department, Hunyuan

Liu Yuhong (刘煜宏) is Vice President of Tencent Cloud and head of Tencent's big data and large model initiatives (Hunyuan). He joined Tencent in 2005 and has over 20 years of experience in the telecom and internet industries. He leads the development of Tencent's big data platforms (from Gen 1 through Gen 4) and the Hunyuan large language model. He oversees Tencent's AI data strategy and infrastructure for the Hunyuan platform. He has extensive experience in cloud computing, big data, AI technologies, and ToB/ToG services. He has been a featured speaker at conferences including the Peking University National Development Institute Greater Bay Area Forum (2024) and the GeekPark interview series.
- Known for: VP Tencent Cloud; Head of Tencent Big Data and Hunyuan LLM; 20+ years at Tencent (since 2005); Built Tencent big data platforms Gen 1-4

### Han Hu (胡瀚)
Head of Multimodal Large Model R&D, Tencent Hunyuan (joined 2024)

Han Hu (胡瀚) is a Distinguished Scientist at Tencent, leading Hunyuan multimodal large model R&D. He earned his B.S. (2008) and Ph.D. (2014) from Tsinghua. He was at Baidu Deep Learning Research Institute (2014-2016) and MSRA Visual Computing Group (2016-2024) before joining Tencent in 2024. His Swin Transformer won the ICCV 2021 Marr Prize.
- Known for: Distinguished Scientist, Tencent Hunyuan multimodal lead (2024); Swin Transformer (ICCV 2021 Marr Prize); MSRA Visual Computing Group (2016-2024)
- Education: Tsinghua University (清华大学)

### WANG Di (王迪)
Deputy GM, LLM Department, Tencent Hunyuan

WANG Di (王迪) is Deputy General Manager of the LLM Department at Tencent and head of the Taiji Machine Learning Platform and Hunyuan large model technology. He joined Tencent in 2008 and has over a decade of deep technical experience in AI, spanning large-scale generative models, sparse models for search/recommendation/advertising, search platforms, GPU computing, and task scheduling. He previously served as Assistant General Manager of TEG-AI Platform Dept. He leads the technical development of Hunyuan's MoE-based models, which have grown from billions to trillions of parameters. He has spoken at CCF conferences and industry forums about trillion-parameter MoE system engineering.
- Known for: Deputy GM LLM Dept, Tencent; Head of Taiji ML Platform and Hunyuan LLM tech; Joined Tencent 2008

### CAN Xu (徐灿)
Senior Researcher, Tencent Hunyuan (joined May 2025)

CAN Xu (徐灿) is a Senior Researcher at Tencent Hunyuan, having joined in May 2025. He is the creator of the WizardLM project. He earned his B.S. from Sichuan University and his M.S. from Peking University (advisor: Yu Jingsong). He joined Microsoft in 2017 as a Senior Research Scientist, where he was a core contributor to Microsoft XiaoIce, Bing, WizardLM, and Phi-3. He led the development of WizardLM 1&2, WizardCoder, and WizardMath. He has published 40+ papers at NeurIPS, ACL, ICLR, EMNLP, AAAAI with 3,300+ Google Scholar citations. In May 2025, he announced on X that he and his team had left Microsoft to join Tencent Hunyuan.
- Known for: Senior Researcher, Tencent Hunyuan (May 2025); Creator of WizardLM; WizardLM 1&2, WizardCoder, WizardMath; 40+ papers, 3,300+ citations; Microsoft XiaoIce, Bing, Phi-3 core contributor

### Qingfeng Sun (孙庆峰)
Senior Researcher, Tencent Hunyuan (joined May 2025)

Qingfeng Sun (孙庆峰) is a Senior Researcher at Tencent Hunyuan, having joined in December 2024. He is a co-founder of the WizardLM project. He earned his B.S. from Hefei University of Technology and his M.S. from Peking University. He was a Microsoft NLP Research Scientist, where he focused on large language model development and contributed to core deep models for Bing and Office 365. He co-created Evol-Instruct, RLEIF, and Arena-Learning methods. He was the first WizardLM team member to join Tencent Hunyuan (Dec 2024), with others following in spring 2025.
- Known for: Senior Researcher, Tencent Hunyuan (Dec 2024); WizardLM co-founder; Evol-Instruct, RLEIF, Arena-Learning co-creator; Former Microsoft NLP Research Scientist (Bing, Office 365)

### Weijie Kong
Tencent Hunyuan video generation tech lead

Weijie Kong is a video generation researcher at Tencent Hunyuan. He is the first author of the HunyuanVideo paper (arXiv 2024), Tencent's systematic framework for large video generative models. The model uses a DiT architecture with 13B parameters and supports both text-to-video and image-to-video generation.
- Known for: Tencent Hunyuan video generation models; HY-OmniWeaving

### SUN Xingwu (孙兴武)
Tencent Hunyuan LLM department researcher

SUN Xingwu (孙兴武) is a Machine Learning Platform Expert at Tencent and the pre-training lead for the Hunyuan large model. He has presented on Hunyuan's technical achievements at industry conferences, including the Open Atom Developer Conference (2024), where he introduced Tencent's self-developed Starlink high-performance computing network, the Angel training framework, and applications of Hunyuan in advertising recommendation, code generation, and AI customer service.
- Known for: Tencent ML Platform Expert; Hunyuan pre-training lead; Starlink high-performance computing network; Angel training framework

### Luo Fuli (罗福莉)
Head of Xiaomi MiMo large-model team (November 2025)

Luo Fuli (罗福莉, born 1995) is the Head of Xiaomi's MiMo large model team, having joined in November 2025. She was born in Yibin, Sichuan, and earned her B.S. in Computer Science from Beijing Normal University and her master's degree from Peking University in computational linguistics. She joined High-Flyer Capital and subsequently DeepSeek in 2022, where she was a core developer on DeepSeek-V2 and other models. She gained widespread media attention as a '95s AI genius' and was reportedly recruited by Xiaomi founder Lei Jun with a multi-million-yuan (roughly high-six-figure USD-equivalent) annual salary. In March 2026, she released MiMo-V2, and in April 2026, MiMo-V2.5 series featuring the 1.02T parameter MiMo-V2.5-Pro model, which reportedly surpassed DeepSeek-V4-Pro and Kimi K2.6 on multiple benchmarks. She is known for requiring her team to have at least 100 AI conversations per day to understand AI's capabilities and limitations.
- Known for: Xiaomi MiMo-V2 (Mar 2026); Xiaomi MiMo-V2.5 series (Apr 2026); DeepSeek-V2 core developer
- Education: Beijing Normal University (北京师范大学), Peking University (北京大学)

### Wang Bin (王斌)
Director, Xiaomi AI Lab; Vice Chairman of Xiaomi Technical Committee

Wang Bin (王斌) is the Director of Xiaomi AI Lab and Vice Chairman of the Xiaomi Technical Committee. He joined Xiaomi in 2018 from the Chinese Academy of Sciences where he was a researcher and Ph.D. advisor for over 20 years, leading information retrieval and NLP teams at the Institute of Computing Technology and Institute of Information Engineering. He holds a Ph.D. from the Institute of Computing Technology, Chinese Academy of Sciences. At Xiaomi, he oversees the company's AI strategy, including the AI Lab which is the core department for Xiaomi's AI initiatives. The AI Lab focuses on NLP, computer vision, speech, and large model research, supporting products like Xiao Ai (小爱同学), IoT, autonomous driving, and robotics. Under his leadership, Xiaomi's large model team was formally established in April 2023 with Luan Jian as the team lead.
- Known for: Director of Xiaomi AI Lab; Vice Chairman of Xiaomi Technical Committee; Led formation of Xiaomi's large model team (2023)
- Education: Institute of Computing Technology, CAS (中科院计算所)

### Luan Jian (栾剑)
(former) Head of Xiaomi AI Lab large-model team (April 2023–2025)

Luan Jian (栾剑) was the head of Xiaomi AI Lab's large model team from April 2023. He is a well-known AI expert who previously served as Chief Speech Scientist at Microsoft Xiaobing (Microsoft XiaoIce), Senior Speech Scientist at Microsoft (China) Engineering Institute, and a researcher at Toshiba (China) Research Institute. He has over 20 years of experience in speech and language technology research, covering voiceprint recognition, speech synthesis, speech recognition, singing synthesis, human-machine dialogue, machine translation, large language models, and multimodal large models. He has published over 100 papers. At Xiaomi, he led the development of MiLM, MiMM, X-Studio (AI singing synthesis), and super-realistic speech synthesis systems. He reported to Wang Bin, Director of Xiaomi AI Lab.
- Known for: Head of Xiaomi AI Lab large model team (Apr 2023-2025); Microsoft Xiaobing Chief Speech Scientist; 100+ published papers; X-Studio AI singing synthesis

### Bingquan Xia (夏炳权)
Tech lead, Xiaomi MiMo LLM-Core

Bingquan Xia (夏炳权) is a tech lead at Xiaomi's MiMo LLM-Core team. He is listed as the first author (core contributor) of the MiMo technical report, which introduced MiMo-7B, a large language model optimized for reasoning tasks. The MiMo-7B-Base model was pre-trained on 25 trillion tokens and the RL-tuned MiMo-7B-RL achieved superior performance on mathematics, code, and general reasoning tasks, surpassing OpenAI o1-mini. He works on the core LLM team at Xiaomi alongside researchers including Bowen Shen, Cici, Dawei Zhu, Di Zhang, and others.
- Known for: MiMo-7B Technical Report (first author/core contributor); MiMo-7B-RL reasoning model

### Wang Jingang (王金刚)
Technology Director, Meituan LongCat large-model team

Wang Jingang (王金刚) is a Meituan algorithm strategy researcher and technology director on the LongCat large-model team. Beijing Daily describes him as a Ph.D. in engineering, senior engineer, CCF senior member, and one of the LongCat R&D leads since 2023. His work connects natural-language processing, local-services knowledge graphs, model compression, retrieval, and Meituan's general-purpose LongCat large-model stack. He has published 40+ papers in CCF-recommended venues and holds 20+ invention patents.
- Known for: Technology Director, Meituan LongCat large-model team; LongCat R&D lead since 2023; Meituan Brain / local-services knowledge graph; 40+ CCF papers and 20+ invention patents

### Pan Xin (潘欣)
Multimodal AI Innovation Lead, Meituan (joined December 2025)

Pan Xin (潘欣) joined Meituan in December 2025 as Multimodal AI Innovation Lead, where 36Kr reports that he leads LongCat App and related multimodal AI applications. He previously worked as a Google Brain / DeepMind researcher on TensorFlow's dynamic graph mode, then returned to China and worked on PaddlePaddle, Tencent's Wuliang deep-learning framework, and ByteDance's AIGC and visual large-model AI platform. His profile is strongest as the systems-to-product bridge in Meituan's LongCat effort.
- Known for: Multimodal AI Innovation Lead, Meituan; LongCat App development lead; TensorFlow dynamic graph mode; ByteDance AIGC and visual large-model AI platform

### Jianing Wang (王嘉宁)
Researcher, Meituan LongCat; first author, LongCat-Flash-Prover

Jianing Wang (王嘉宁) is a Meituan LongCat researcher and first author of LongCat-Flash-Prover, a 560B open-source MoE model for native formal reasoning in Lean4. The Prover report introduces agentic tool-integrated reinforcement learning and reports strong open-weights performance on auto-formalization and theorem-proving benchmarks. Jianing Wang is also listed on R-HORIZON, a benchmark and training-data method for long-horizon reasoning.
- Known for: First author, LongCat-Flash-Prover; 560B MoE formal reasoning model; R-HORIZON reasoning benchmark

### Xunliang Cai (蔡勋梁)
Researcher, Meituan LongCat video direction

Xunliang Cai (蔡勋梁) is a Meituan LongCat researcher who is listed near the front of the LongCat-Video technical report and appears across LongCat-Flash-Prover, R-HORIZON, and LongCat-Next. His visible work makes him the clearest public LongCat card for video generation and native multimodal modeling, rather than another generic author-list entry.
- Known for: LongCat-Video technical report; LongCat-Next native multimodal model; LongCat-Flash-Prover; R-HORIZON reasoning benchmark

### Pang Ruoming (庞若鸣)
Researcher, OpenAI (since February 2026)

Pang Ruoming (庞若鸣, also known as Ruoming Pang) joined OpenAI in February 2026. He earned his B.S. from Shanghai Jiao Tong University (1998), M.S. from USC (2000), and Ph.D. from Princeton (2006). He spent 15+ years at Google leading Zanzibar, Babelfish/Lingvo, and Tacotron 2. He was Apple's AI/ML foundation models lead (2021-2025), then Meta MSL's AI infrastructure lead (2025), before joining OpenAI.
- Known for: OpenAI researcher (Feb 2026); Meta MSL AI infra lead (2025); Apple AI/ML foundation models lead (2021-2025); Google: Zanzibar, Babelfish, Tacotron 2 (15+ years)
- Education: Shanghai Jiao Tong University (上海交通大学), University of Southern California, Princeton University

### Zhao Shengjia (赵晟佳)
Chief Scientist, Meta Superintelligence Labs (since July 25, 2025)

Zhao Shengjia (赵晟佳) is the Chief Scientist at Meta Superintelligence Labs (MSL), appointed on July 25, 2025. He earned his B.S. in Computer Science from Tsinghua University and his Ph.D. from Stanford in 2022. At OpenAI, he was a core contributor to ChatGPT, GPT-4, and o1 reasoning models. He is one of two Chief Scientists at MSL alongside Yann LeCun. He received the ICLR 2022 Outstanding Paper Award.
- Known for: Chief Scientist, Meta MSL (Jul 2025); ChatGPT/GPT-4/o1 contributor (OpenAI); ICLR 2022 Outstanding Paper

### Bi Shuchao (毕树超)
Researcher, Meta Superintelligence Labs (since 2025)

Bi Shuchao (毕树超) joined Meta Superintelligence Labs in 2025. He earned his B.S. in Mathematics from Zhejiang University and Ph.D. from UC Berkeley. He was a former Google Engineering Director and co-founder of YouTube Shorts. At OpenAI, he was manager of multimodal post-training and co-creator of GPT-4o voice mode and o4-mini.
- Known for: Meta MSL researcher (2025); YouTube Shorts co-founder (Google); GPT-4o voice mode co-creator (OpenAI)
- Education: Zhejiang University (浙江大学), UC Berkeley

### Huiwen Chang (常惠雯)
Researcher, Meta Superintelligence Labs (since 2025)

Huiwen Chang (常惠雯) joined Meta Superintelligence Labs in 2025. She is a Tsinghua Yao Class graduate and earned her Ph.D. from Princeton in 2018 under Prof. Adam Finkelstein. She invented MaskGIT and Muse at Google and led GPT-4o image generation at OpenAI.
- Known for: Meta MSL researcher (2025); Inventor of MaskGIT and Muse (Google); GPT-4o image generation lead (OpenAI)
- Education: Tsinghua University (清华大学), Princeton University

### Ji Lin (林骥)
Researcher, Meta Superintelligence Labs (since 2025)

Ji Lin (林骥) joined Meta Superintelligence Labs in 2025. He earned his B.S. from Tsinghua and M.S./Ph.D. from MIT. He was a key contributor to o3, o4-mini, GPT-4o, GPT-4.1, and the Operator reasoning stack at OpenAI. He previously worked at Adobe Research, OmniML, and NVIDIA Research.
- Known for: Meta MSL researcher (2025); OpenAI: o3, o4-mini, GPT-4o, Operator reasoning

### Ren Hongyu (任泓宇)
Researcher, Meta Superintelligence Labs (since 2025)

Ren Hongyu (任泓宇) joined Meta Superintelligence Labs in June 2025. He earned his B.S. from Peking University and Ph.D. from Stanford. At OpenAI, he was post-training lead for o3, o4-mini, o1-mini, and GPT-4o. He received the Apple PhD Fellowship, Baidu PhD Fellowship, and Masayoshi Son Foundation Scholarship.
- Known for: Meta MSL researcher (Jun 2025); OpenAI post-training lead: o3, o4-mini, o1-mini, GPT-4o
- Education: Peking University (北京大学), Stanford University

### Sun Pei (孙培)
Researcher, Meta Superintelligence Labs (since 2025)

Sun Pei (孙培) joined Meta Superintelligence Labs in 2025. He is a Tsinghua alumnus and former Google DeepMind senior researcher on Gemini and Waymo. He is also an Associate Professor at Tsinghua Psychology Department and Chief Researcher at the Tsinghua Brain and Intelligence Laboratory.
- Known for: Meta MSL researcher (2025); Google DeepMind: Gemini, Waymo; Tsinghua Assoc Prof (Psychology)
- Education: Tsinghua University (清华大学), Carnegie Mellon University

### Yu Jiahui (余家辉)
Researcher, Meta Superintelligence Labs (since 2025)

Yu Jiahui (余家辉) joined Meta Superintelligence Labs in 2025. He earned his B.S. from USTC Special Class for the Gifted Young and Ph.D. from UIUC. He worked at Baidu, NVIDIA, and Google DeepMind (Gemini multimodal co-lead) before joining OpenAI in Oct 2023, where he led the Perception team and contributed to o3, o4-mini, GPT-4.1, and GPT-4o. At Meta, he leads multimedia content generation R&D.
- Known for: Meta MSL researcher (2025); OpenAI Perception team lead; Gemini multimodal co-lead (DeepMind)
- Education: University of Science and Technology of China (中国科学技术大学), University of Illinois Urbana-Champaign

### Yuanzhi Li (李远志)
Researcher, Meta Superintelligence Labs (since July 2025)

Yuanzhi Li (李远志) joined Meta Superintelligence Labs in July 2025. He earned his Ph.D. from Princeton in 2018. He was a core contributor to Microsoft's Phi series (Phi-1.5, Phi-3, Phi-4) of small language models and previously worked at OpenAI.
- Known for: Meta MSL researcher (Jul 2025); Microsoft Phi series core contributor; Former OpenAI
- Education: Princeton University

### Yao Shunyu (Stanford physics) (姚顺宇)
Senior Research Scientist, Google DeepMind core Gemini team (since September 29, 2025)

Yao Shunyu (姚顺宇, born ~1997) is a Senior Research Scientist at Google DeepMind, joining on September 29, 2025. He is a theoretical physicist turned AI researcher. He entered Tsinghua University's Physics Department in 2015, where he won the university's highest undergraduate honor (Special Prize Scholarship) and published groundbreaking work on the Non-Hermitian Skin Effect in Physical Review Letters. He earned his Ph.D. from Stanford Institute for Theoretical Physics under Douglas Stanford and Stephen Shenker, then did a postdoc at UC Berkeley before pivoting to AI. He joined Anthropic on October 1, 2024, where he contributed to Claude 3.7 and Claude 4.5 development through reinforcement learning research. He left Anthropic on September 19, 2025, and joined Google DeepMind's Gemini team. He is one of two prominent AI researchers named Yao Shunyu from Tsinghua's class of 2015/2019 - the other (姚顺雨) is now at Tencent.
- Known for: Non-Hermitian Skin Effect discovery (PRL, ~1000 citations); Scramblon theory in quantum chaos; Claude 3.7 and 4.5 RL contributions at Anthropic; Gemini 3 Deep Think at Google DeepMind
- Education: Tsinghua University (清华大学), Stanford University

### Binyuan Hui (惠彬原)
Now at Meta (since January 2026)

Binyuan Hui (惠彬原, born 1999) is a Senior Research Scientist at Meta, having joined in January 2026. He was a child prodigy who entered university at age 15. He earned his M.S. from Tianjin University. He joined Alibaba DAMO Academy's Intelligent Dialogue Team as an intern, then became a full-time employee after graduation. At Alibaba's Tongyi Lab, he served as algorithm expert and lead of Qwen Code (Qwen series model code training). He was a core contributor to Qwen series papers and participated in the 'Hundred Famous Enterprises Entering Campus' lecture series. He left Alibaba in January 2026 and joined Meta. His research focuses on NLP, semantic parsing, intelligent dialogue, and large-scale pre-trained models. He published the paper 'STAR: SQL Guided Pre-Training for Context-dependent Text-to-SQL Parsing' at EMNLP 2022 and contributed to TableQA technology that topped multiple leaderboards.
- Known for: Meta Senior Research Scientist (Jan 2026); Qwen Code lead, Alibaba Tongyi Lab; STAR paper (EMNLP 2022); TableQA technology

### Andrew Ng (吴恩达)
Stanford CS (founder DeepLearning.AI, AI Fund, Landing AI)

Andrew Ng (吴恩达) is a globally recognized leader in AI and machine learning. He is the Founder of DeepLearning.AI, Managing General Partner at AI Fund, Executive Chairman of LandingAI, and Chairman and Co-Founder of Coursera. He is also an Adjunct Professor of Computer Science at Stanford University. Ng previously served as Chief Scientist at Baidu and as the founding lead of the Google Brain team. He has authored or co-authored over 200 research papers in machine learning, robotics, and related fields. He holds a B.Sc. from Carnegie Mellon University, an M.Sc. from MIT, and a Ph.D. from UC Berkeley. Named to Time magazine's 100 Most Influential People in 2012 and the Time100 AI list in 2023, Ng has taught over 8 million students through his online AI courses.
- Known for: Google Brain (founding lead); Coursera (co-founder); DeepLearning.AI (founder); AI Fund (founding partner); LandingAI (Executive Chairman); Deep Learning Specialization (course)
- Education: Carnegie Mellon University, Massachusetts Institute of Technology, UC Berkeley

### Harry Shum (沈向洋)
Founder, IDEA Research (Beijing)

Harry Shum (沈向洋, born October 1966) is the Founding Chairman of the International Digital Economy Academy (IDEA, 粤港澳大湾区数字经济研究院) in Shenzhen and an Adjunct Professor at Tsinghua University's Institute for Advanced Study. He previously served as Executive Vice President of Microsoft's AI and Research group (2013-2020), overseeing Microsoft's global AI strategy and research labs. Shum was a founding member of Microsoft Research China (later renamed Microsoft Research Asia) in 1998 and served as its Managing Director from 2004 to 2007. He earned his B.S. from Southeast University, MPhil from the University of Hong Kong (1991), and Ph.D. in Robotics from Carnegie Mellon University (1996) under Raj Reddy. He is a Fellow of IEEE (2006) and ACM (2007), and a foreign member of both the US National Academy of Engineering (2017) and the UK's Royal Academy of Engineering (2018). He has published over 200 papers and holds 50+ US patents.
- Known for: IDEA Research (Founding Chairman, 2020-present); Microsoft Research Asia (founding member, former MD); Microsoft EVP of AI and Research (2013-2020); Microsoft Bing (VP product development, 2007-2013)

### Song Han (韩松)
Associate Professor MIT EECS

Song Han (韩松) is an Associate Professor in the Electrical Engineering and Computer Science (EECS) department at MIT. He is an expert in deep learning and computer architecture. Song Han earned his B.S. from Tsinghua University and his Ph.D. from Stanford University. He proposed the widely used 'Deep Compression' technique for efficient AI computing, and pioneered EIE (Efficient Inference Engine) which introduced weight sparsity to modern AI chips, influencing NVIDIA's Ampere GPU architecture. He also founded the TinyML research field, bringing machine learning to IoT devices and enabling edge machine learning. In February 2023, he received the Sloan Research Fellowship.
- Known for: Deep Compression technique for efficient AI computing; EIE (Efficient Inference Engine); TinyML research field founder; Influenced NVIDIA Ampere GPU architecture
- Education: Tsinghua University (清华大学), Stanford University


# Investors

## Alibaba Group (阿里巴巴)

Hangzhou internet, cloud, commerce, and AI platform company. Its AI stack spans Alibaba Cloud, Tongyi/Qwen models, enterprise AI services, and the broader Taobao/Tmall commerce ecosystem.

- **Type:** corporate
- **Location:** Hangzhou (杭州)
- **Tracked lab investments:** Zhipu AI / Z.ai (智谱AI); Moonshot AI / Kimi (月之暗面); MiniMax (稀宇科技); 01.AI (零一万物); Baichuan AI (百川智能)
- **Site:** https://www.alibabagroup.com/en-US

## Tencent (腾讯)

Shenzhen-based internet, games, cloud, payments, and social-platform group. Its internal AI work centers on Hunyuan, with distribution through WeChat, Tencent Cloud, games, ads, and enterprise services.

- **Type:** corporate
- **Location:** Shenzhen (深圳)
- **Tracked lab investments:** Zhipu AI / Z.ai (智谱AI); Moonshot AI / Kimi (月之暗面); Baichuan AI (百川智能); StepFun (阶跃星辰); Sogou (搜狗)
- **Site:** https://www.tencent.com/en-us/

## ByteDance (字节跳动)

Private global internet and AI platform company behind TikTok, Douyin, Toutiao, CapCut, and Doubao. Its Seed research organization develops foundation models for ByteDance products and APIs.

- **Type:** corporate
- **Location:** Beijing (北京)
- **Site:** https://www.bytedance.com/en/

## HongShan (formerly Sequoia China) (红杉中国)

Formerly Sequoia Capital China; a venture and private-equity platform founded in 2005 and led by Neil Shen. Invests across technology, healthcare, and consumer companies, with more than $55B in assets under management.

- **Type:** vc
- **Tracked lab investments:** Moonshot AI / Kimi (月之暗面); Zhipu AI / Z.ai (智谱AI); MiniMax (稀宇科技)
- **Site:** https://www.hsgcap.com/about-us/

## Hillhouse Capital (高瓴资本)

Long-horizon global investment manager founded by Zhang Lei in 2005. Invests across public equity, private equity, real assets, and credit, with a focus on healthcare, consumer, industrials, business services, and energy transition.

- **Type:** pe
- **Tracked lab investments:** MiniMax (稀宇科技)
- **Site:** https://www.hillhouseinvestment.com/

## Qiming Venture Partners (启明创投)

China venture firm founded in 2006, investing in early and growth-stage technology/consumer and healthcare companies. Manages US dollar and RMB funds with about $9.5B in capital raised.

- **Type:** vc
- **Tracked lab investments:** StepFun (阶跃星辰)
- **Site:** https://www.qimingvc.com/en/about

## Sinovation Ventures (创新工场)

Chinese technology venture firm founded in 2009 by Kai-Fu Lee. Manages USD and RMB funds, backs early-stage technology and AI companies, and incubated 01.AI.

- **Type:** vc
- **Tracked lab investments:** 01.AI (零一万物)
- **Site:** https://www.sinovationventures.com/

## miHoYo (米哈游)

Shanghai game developer and publisher behind Genshin Impact and Honkai. The company is known for anime-style IP, global live-service games, and unusually strong in-house technology and art pipelines.

- **Type:** corporate
- **Location:** Shanghai (上海)
- **Tracked lab investments:** MiniMax (稀宇科技)
- **Site:** https://www.mihoyo.com/en

## IDG Capital (IDG资本)

One of the earliest foreign-backed venture investors in China, active since 1993. Invests from seed through venture, growth, buyout, and public-market stages across consumer and technology sectors.

- **Type:** vc
- **Tracked lab investments:** Moonshot AI / Kimi (月之暗面)
- **Site:** https://www.idgcapital.com/

## Xiaomi (小米)

Beijing consumer electronics, IoT, EV, and internet-services company built around phones, smart-home hardware, software services, and the Xiaomi SU7 auto push. Its internal large-model effort is Xiaomi MiMo.

- **Type:** corporate
- **Location:** Beijing (北京)
- **Tracked lab investments:** Zhipu AI / Z.ai (智谱AI); 01.AI (零一万物); Baichuan AI (百川智能)
- **Site:** https://www.mi.com/global/about/

## 5Y Capital (五源资本)

Shanghai venture capital firm formerly known as Morningside Venture Capital. Founded in 2008, it manages over $5B and invests in technology, life sciences, and consumer innovation; earlier portfolio names include Xiaomi, Kuaishou, XPeng, Pony.ai, and SenseTime.

- **Type:** vc
- **Tracked lab investments:** StepFun (阶跃星辰)
- **Site:** https://www.5ycap.com/en/about/

## Meituan (美团)

Chinese local-services and delivery platform spanning food delivery, in-store services, hotel/travel, retail, logistics, and merchant software. It is one of the country's largest consumer internet operating systems.

- **Type:** corporate
- **Location:** Beijing (北京)
- **Tracked lab investments:** Zhipu AI / Z.ai (智谱AI)
- **Site:** https://about.meituan.com/en

## Ant Group (蚂蚁集团)

Alibaba-affiliated fintech and technology group operating Alipay, digital payments, credit-tech, wealth-tech, insurance-tech, and enterprise blockchain/cloud services.

- **Type:** corporate
- **Location:** Hangzhou (杭州)
- **Tracked lab investments:** Zhipu AI / Z.ai (智谱AI)
- **Site:** https://www.antgroup.com/en

## Prosperity7 (Saudi Aramco)

Aramco Ventures' $1B global VC fund for disruptive technologies and business models outside the energy sector, with offices including Dhahran, Palo Alto, New York, Beijing, and Shanghai.

- **Type:** sovereign
- **Tracked lab investments:** Zhipu AI / Z.ai (智谱AI)
- **Site:** https://aramcoventures.com/

## China Big Fund (国家集成电路产业投资基金)

China's state-backed National Integrated Circuit Industry Investment Fund, created in 2014 to support the domestic semiconductor supply chain. Its mandate covers chip design, fabrication, equipment, materials, and strategic semiconductor capacity.

- **Type:** state
- **Tracked lab investments:** DeepSeek (深度求索)

## Pudong Venture Capital (浦东创投)

Shanghai Pudong venture and industrial-capital platform. Part of Pudong's state-backed technology-finance system, with vehicles spanning angel, science-and-technology, industrial, and strategic investment.

- **Type:** state
- **Location:** Shanghai (上海)
- **Tracked lab investments:** StepFun (阶跃星辰); Zhipu AI / Z.ai (智谱AI)
- **Site:** https://www.pdvc.com/

## Zhangjiang Group (张江集团)

Shanghai state-backed developer and operator of Zhangjiang Science City, founded in 1992. Its role combines science-park development, industry cultivation, technology services, and investment in the Zhangjiang innovation ecosystem.

- **Type:** state
- **Location:** Shanghai (上海)
- **Tracked lab investments:** Zhipu AI / Z.ai (智谱AI)
- **Site:** https://www.zjpark.com/en/guanyu.html

## Shanghai SSCIC / Shanghai Guotou (Fortera) (上海国投 / 上海国际集团)

Shanghai State-owned Capital Investment platform focused on fund management and innovation incubation. It positions itself as patient state capital for early, small, long-term, hard-tech investment.

- **Type:** state
- **Location:** Shanghai (上海)
- **Tracked lab investments:** StepFun (阶跃星辰)
- **Site:** https://www.sh-ssci.com/

## China Life Private Equity (国寿资本)

Professional private-equity platform under China Life, established in 2016 in the Shanghai Free Trade Zone. Known for insurance-capital PE funds, especially healthcare and hard-tech related strategies.

- **Type:** state
- **Tracked lab investments:** StepFun (阶跃星辰)
- **Site:** https://www.chinalifepe.com/en/

## Beijing AI Industry Investment Fund (北京人工智能产业基金)

Beijing municipal/state-backed AI industry fund used to channel patient capital into local AI and strategic-technology companies. It fits the city's policy toolkit around model companies, computing power, datasets, and Zhongguancun commercialization.

- **Type:** state
- **Location:** Beijing (北京)
- **Tracked lab investments:** Baichuan AI (百川智能)

## Huajin Capital (华金资本)

Zhuhai-listed investment platform (000532) combining equity investment management, industrial operations, and science-park/innovation services. Focus areas include advanced manufacturing, healthcare, emerging technology, new energy, and environmental protection.

- **Type:** state
- **Tracked lab investments:** DeepSeek (深度求索)
- **Site:** https://huajinct.com/about/Team.aspx?nav=0&p=0

## Zhejiang Oriental (浙江东方)

Hangzhou-based, SSE-listed financial holding group offering trust, futures, insurance, leasing, fund investment, wealth management, and asset-management services. It sits inside Zhejiang's provincial financial and industrial-capital system.

- **Type:** state
- **Location:** Hangzhou (杭州)
- **Tracked lab investments:** DeepSeek (深度求索)
- **Site:** https://www.zjorient.com/

## Gaorong Capital (高榕资本)

Beijing venture firm focused on early and growth-stage TMT, AI, and frontier-technology investments. Formerly known internationally as Banyan Capital.

- **Type:** vc
- **Tracked lab investments:** Moonshot AI / Kimi (月之暗面)
- **Site:** https://www.gaorongvc.com/

## Shenzhen Capital Group (深创投)

Shenzhen government-established venture capital group founded in 1999. State-owned platform focused on strategic and emerging industries, especially innovative high-tech companies and national industrial champions.

- **Type:** state
- **Location:** Shenzhen (深圳)
- **Tracked lab investments:** Baichuan AI (百川智能)
- **Site:** https://www.szvc.com.cn/


# Universities

## Tsinghua University (清华大学)

Single biggest source of China AI talent. Spawned Zhipu, Moonshot, Megvii. Major sub-units: KEG (Knowledge Engineering Group), Yao Class / IIIS, AIR.

- **Location:** Beijing (北京)
- **Labs that originated here:** Zhipu AI / Z.ai (智谱AI)
- **Tracked researchers:** Zhihong Shao (邵智宏); Zhibin Gou (苟志斌); Tang Jie (唐杰); Li Juanzi (李娟子); Zhang Peng (张鹏); Zhengxiao Du (杜正晓); Ming Ding (丁铭); Yang Zhilin (杨植麟); Zhang Yutao (张宇韬); Zhou Xinyu (周昕宇); Wu Yuxin (吴育昕); Yan Junjie (闫俊杰); Heng Ji (季姮); Bei Chen (陈蓓); Wang Xiaochuan (王小川); Ru Liyun (茹立云); Zhu Yibo (朱亦博); Yin Qi (印奇); Yao Shunyu (姚顺雨); Han Hu (胡瀚); Huiwen Chang (常惠雯); Sun Pei (孙培); Yao Shunyu (Stanford physics) (姚顺宇); Song Han (韩松)

## Peking University (北京大学)

Core team contributors at DeepSeek (Damai Dai, Qihao Zhu); Lin Junyang's alma mater (Qwen).

- **Location:** Beijing (北京)
- **Tracked researchers:** Deli Chen (陈德立); Damai Dai (代达劢); Qihao Zhu (朱启昊); Zhao Pengyu (赵鹏宇); Wenhao Huang (黄文灏); Lin Junyang (林俊旸); Zhou Chang (周畅); Luo Fuli (罗福莉); Ren Hongyu (任泓宇)

## Beihang University (北京航空航天大学)

'Seven Sons of National Defense' university. Major DeepSeek talent source: Yu Wu, Zhihong Shao undergrad, Dejian Yang lab.

- **Location:** Beijing (北京)
- **Tracked researchers:** Yu Wu (吴俣); Zhihong Shao (邵智宏); Dejian Yang (杨德建); Zhou Yucong (周彧聪)

## Zhejiang University (浙江大学)

Liang Wenfeng's alma mater (DeepSeek). Symbolic of Hangzhou AI.

- **Location:** Hangzhou (杭州)
- **Tracked researchers:** Liang Wenfeng (梁文锋); Liang Wenfeng (梁文锋); Junxiao Song (宋俊潇); Bi Shuchao (毕树超)

## Shanghai Jiao Tong University (上海交通大学)

Anchor of Shanghai AI talent pipeline; Pang Ruoming's undergrad alma mater.

- **Location:** Shanghai (上海)
- **Tracked researchers:** Yongyi Hu (胡咏毅); Pang Ruoming (庞若鸣)

## University of Science and Technology of China (中国科学技术大学)

Hefei-based elite STEM university. Zhou Jingren, Jiao Binxing, Yu Jiahui alma mater.

- **Tracked researchers:** Yang Bin (杨斌); Jiao Binxing (焦斌星); Zhou Jingren (周靖人); Jiashi Feng (冯佳时); Yu Jiahui (余家辉)

## Harbin Institute of Technology (哈尔滨工业大学)

'Seven Sons of National Defense'. Chen Weipeng (Baichuan) alma mater.

- **Tracked researchers:** Bei Chen (陈蓓); Chen Weipeng (陈未鹏)

## Sun Yat-sen University (中山大学)

Daya Guo's alma mater (DeepSeek R1 first author).

- **Tracked researchers:** Daya Guo (郭达雅); Su Jianlin (苏剑林)

## Beijing University of Posts and Telecommunications (北京邮电大学)

Zhibin Gou's undergrad alma mater.

- **Tracked researchers:** Zhibin Gou (苟志斌)

## Shanxi University (山西大学)

Li Juanzi's BS + MS alma mater.

- **Tracked researchers:** Li Juanzi (李娟子); Li Juanzi (李娟子)

## Southeast University (东南大学)

Yan Junjie's undergrad alma mater.

- **Tracked researchers:** Yan Junjie (闫俊杰)

## Beijing Normal University (北京师范大学)

Luo Fuli's undergrad alma mater.

- **Tracked researchers:** Luo Fuli (罗福莉)

## Central South University (中南大学)

Bowen Yu's undergrad alma mater.

- **Tracked researchers:** Bowen Yu (郁博文)

## Fudan University (复旦大学)

Zhou Chang's undergrad alma mater (ByteDance Seed multimodal lead).

- **Location:** Shanghai (上海)
- **Tracked researchers:** Zhou Chang (周畅)

## Institute of Automation, CAS (中科院自动化所)

Chinese Academy of Sciences research institute. Yan Junjie (MiniMax) PhD; Wang Bingning (Baichuan) PhD.

- **Location:** Beijing (北京)
- **Tracked researchers:** Yan Junjie (闫俊杰); Yang Bin (杨斌); Wang Bingning (王炳宁)

## Institute of Computing Technology, CAS (中科院计算所)

Chinese Academy of Sciences institute. Wang Bin (Xiaomi AI Lab) PhD.

- **Location:** Beijing (北京)
- **Tracked researchers:** Wang Bin (王斌)

## Institute of Information Engineering, CAS (中科院信息工程研究所)

Chinese Academy of Sciences institute. Bowen Yu PhD.

- **Location:** Beijing (北京)
- **Tracked researchers:** Bowen Yu (郁博文)

## Nanjing University (南京大学)

Wu Yonghui's (ByteDance Seed) undergrad alma mater.

- **Tracked researchers:** Wu Yonghui (吴永辉)

## Carnegie Mellon University

Yang Zhilin PhD (Salakhutdinov + Cohen). Wu Yuxin MS. Sun Pei MS.

- **Tracked researchers:** Yang Zhilin (杨植麟); Wu Yuxin (吴育昕); Kai-Fu Lee (李开复); Sun Pei (孙培); Andrew Ng (吴恩达)

## Princeton University

Yao Shunyu (Tencent) PhD. Pang Ruoming PhD.

- **Tracked researchers:** Yao Shunyu (姚顺雨); Pang Ruoming (庞若鸣); Huiwen Chang (常惠雯); Yuanzhi Li (李远志)

## Columbia University

Zhou Jingren PhD. Yin Qi MS.

- **Tracked researchers:** Kai-Fu Lee (李开复); Yin Qi (印奇); Zhou Jingren (周靖人)

## UC Santa Barbara

Zhu Yibo (StepFun CTO) PhD.

- **Tracked researchers:** Zhu Yibo (朱亦博)

## UC Riverside

Wu Yonghui MS + PhD.

- **Tracked researchers:** Wu Yonghui (吴永辉)

## University of Illinois Urbana-Champaign

Yu Jiahui PhD. Heng Ji full professor.

- **Tracked researchers:** Yu Jiahui (余家辉)

## New York University

Heng Ji PhD.

- **Tracked researchers:** Heng Ji (季姮)

## UC Berkeley

Bi Shuchao grad.

- **Tracked researchers:** Bi Shuchao (毕树超); Andrew Ng (吴恩达)

## Johns Hopkins University

Yun Yeyi (MiniMax COO) BS.

- **Tracked researchers:** Yun Yeyi (贠烨祎)

## Hong Kong University of Science and Technology

Junxiao Song (DeepSeek, GRPO inventor) PhD.

- **Tracked researchers:** Junxiao Song (宋俊潇)

## Cornell University

Zhengxiao Du (Zhipu) research intern.


## University of Notre Dame

Yuxiao Dong (Zhipu) PhD.

- **Tracked researchers:** Yuxiao Dong (董昱潇)

## Stanford University

Ren Hongyu PhD. Yao Shunyu (physics) PhD. Andrew Ng + Fei-Fei Li faculty.

- **Tracked researchers:** Ren Hongyu (任泓宇); Yao Shunyu (Stanford physics) (姚顺宇); Song Han (韩松)

## Massachusetts Institute of Technology

Song Han Associate Professor (EECS).

- **Tracked researchers:** Andrew Ng (吴恩达)

## University at Buffalo (SUNY)

Jiang Daxin (StepFun CEO) PhD.

- **Tracked researchers:** Jiang Daxin (姜大昕)

## University of Toronto

Yang Bin PhD under Raquel Urtasun.

- **Tracked researchers:** Yang Bin (杨斌)

## National University of Singapore

Jiashi Feng PhD.

- **Tracked researchers:** Jiashi Feng (冯佳时)

## University of Southern California

Pang Ruoming MS.

- **Tracked researchers:** Pang Ruoming (庞若鸣)

## École Polytechnique

Yongyi Hu (ex-MiniMax) joint French engineering program.

- **Tracked researchers:** Yongyi Hu (胡咏毅)


# Models

## Claude Fable 5.1

- **Lab:** Anthropic
- **Artificial Analysis Intelligence Index:** 65.7
- **Value score:** 3.3
- **Input price:** $10 per 1M tokens
- **Output price:** $50 per 1M tokens
- **Context:** 1000K tokens
- **Modality:** text, image, file
- **Weights:** closed

## Claude Opus 5

- **Lab:** Anthropic
- **Artificial Analysis Intelligence Index:** 63.1
- **Value score:** 6.3
- **Input price:** $5 per 1M tokens
- **Output price:** $25 per 1M tokens
- **Context:** 1000K tokens
- **Modality:** text, image, file
- **Weights:** closed

## GPT-5.6 Sol

- **Lab:** OpenAI
- **Artificial Analysis Intelligence Index:** 60.9
- **Value score:** 15.2
- **Input price:** $2 per 1M tokens
- **Output price:** $10 per 1M tokens
- **Context:** 1050K tokens
- **Modality:** file, image, text
- **Weights:** closed

## Kimi K3

- **Lab:** Moonshot
- **Artificial Analysis Intelligence Index:** 59.7
- **Value score:** 10
- **Input price:** $3 per 1M tokens
- **Output price:** $15 per 1M tokens
- **Context:** 1049K tokens
- **Modality:** text, image, video
- **Weights:** open

## GLM-5.3

- **Lab:** Zhipu
- **Artificial Analysis Intelligence Index:** 59.5
- **Value score:** 27.7
- **Input price:** $1.4 per 1M tokens
- **Output price:** $4.4 per 1M tokens
- **Context:** 1311K tokens
- **Modality:** text
- **Weights:** closed

## Gemini 3.8 Flash

- **Lab:** Google
- **Artificial Analysis Intelligence Index:** 58.7
- **Value score:** 39.1
- **Input price:** $0.75 per 1M tokens
- **Output price:** $3.75 per 1M tokens
- **Context:** 1049K tokens
- **Modality:** text, image, video, file, audio
- **Weights:** closed

## GPT-5.6 Terra

- **Lab:** OpenAI
- **Artificial Analysis Intelligence Index:** 56.6
- **Value score:** 12.6
- **Input price:** $2 per 1M tokens
- **Output price:** $12 per 1M tokens
- **Context:** 1050K tokens
- **Modality:** file, image, text
- **Weights:** closed

## GPT-5.6 Luna

- **Lab:** OpenAI
- **Artificial Analysis Intelligence Index:** 52.3
- **Value score:** 116.2
- **Input price:** $0.2 per 1M tokens
- **Output price:** $1.2 per 1M tokens
- **Context:** 1050K tokens
- **Modality:** file, image, text
- **Weights:** closed

## Gemini 3.1 Pro

- **Lab:** Google
- **Artificial Analysis Intelligence Index:** 47.7
- **Value score:** 10.6
- **Input price:** $2 per 1M tokens
- **Output price:** $12 per 1M tokens
- **Context:** 1049K tokens
- **Modality:** audio, file, image, text, video
- **Weights:** closed

## MiniMax M3

- **Lab:** MiniMax
- **Artificial Analysis Intelligence Index:** 45.4
- **Value score:** 86.5
- **Input price:** $0.3 per 1M tokens
- **Output price:** $1.2 per 1M tokens
- **Context:** 1049K tokens
- **Modality:** text, image, video
- **Weights:** open

## DeepSeek V4 Pro

- **Lab:** DeepSeek
- **Artificial Analysis Intelligence Index:** 45.3
- **Value score:** 35
- **Input price:** $1.034 per 1M tokens
- **Output price:** $2.069 per 1M tokens
- **Context:** 1049K tokens
- **Modality:** text
- **Weights:** open

## Kimi K2.6

- **Lab:** Moonshot
- **Artificial Analysis Intelligence Index:** 45.1
- **Value score:** 26.3
- **Input price:** $0.95 per 1M tokens
- **Output price:** $4 per 1M tokens
- **Context:** 262K tokens
- **Modality:** text, image
- **Weights:** open

## MiMo V2.5 Pro

- **Lab:** Xiaomi
- **Artificial Analysis Intelligence Index:** 42.9
- **Value score:** 78.9
- **Input price:** $0.435 per 1M tokens
- **Output price:** $0.87 per 1M tokens
- **Context:** 1050K tokens
- **Modality:** text
- **Weights:** open

## DeepSeek V4 Flash

- **Lab:** DeepSeek
- **Artificial Analysis Intelligence Index:** 42.1
- **Value score:** 416.8
- **Input price:** $0.081 per 1M tokens
- **Output price:** $0.162 per 1M tokens
- **Context:** 1049K tokens
- **Modality:** text
- **Weights:** open

## Qwen3.6 Plus

- **Lab:** Alibaba / Qwen
- **Artificial Analysis Intelligence Index:** 40.5
- **Value score:** 55.4
- **Input price:** $0.325 per 1M tokens
- **Output price:** $1.95 per 1M tokens
- **Context:** 1000K tokens
- **Modality:** text, image, video
- **Weights:** open

## Grok 4.3

- **Lab:** xAI
- **Artificial Analysis Intelligence Index:** 37.9
- **Value score:** 24.3
- **Input price:** $1.25 per 1M tokens
- **Output price:** $2.5 per 1M tokens
- **Context:** 1000K tokens
- **Modality:** text, image, file
- **Weights:** closed

## Step 3.7 Flash

- **Lab:** StepFun
- **Artificial Analysis Intelligence Index:** 30.9
- **Value score:** 70.5
- **Input price:** $0.2 per 1M tokens
- **Output price:** $1.15 per 1M tokens
- **Context:** 262K tokens
- **Modality:** text, image, video
- **Weights:** open

## Mistral Medium 3.5

- **Lab:** Mistral
- **Artificial Analysis Intelligence Index:** 30.4
- **Value score:** 10.1
- **Input price:** $1.5 per 1M tokens
- **Output price:** $7.5 per 1M tokens
- **Context:** 262K tokens
- **Modality:** text, image, file
- **Weights:** closed

## Llama 4 Maverick

- **Lab:** Meta
- **Artificial Analysis Intelligence Index:** 14.5
- **Value score:** 44.8
- **Input price:** $0.2 per 1M tokens
- **Output price:** $0.696 per 1M tokens
- **Context:** 1049K tokens
- **Modality:** text, image
- **Weights:** open

## Hunyuan A13B

- **Lab:** Tencent
- **Input price:** $0.14 per 1M tokens
- **Output price:** $0.57 per 1M tokens
- **Context:** 131K tokens
- **Modality:** text
- **Weights:** open