Hongyu Ren is a research scientist known for his work on large language models, particularly the o-series models and GPT-4o at OpenAI, and his subsequent move to Meta's Superintelligence Labs.

Education

Hongyu Ren earned a Bachelor of Science degree with honors in Computer Science from Peking University. He later pursued graduate studies at Stanford University, where he completed a Ph.D. in Computer Science. ^{[1] [2] [5]}

Career

Hongyu Ren is a computer scientist with experience in machine learning research across both academic and industry settings. Since July 2023, he has worked as a Research Scientist at OpenAI, contributing to the development of several language models. His work has included involvement in the design and implementation of OpenAI’s o1, o1-mini, o3-mini, and GPT-4o models. He has also led efforts in the post-training phase of model development.

Prior to OpenAI, Ren held research positions at multiple technology companies. In 2021, he completed two internships at Apple, where he worked on natural language processing tasks related to question answering in voice assistant systems. From 2020 to 2021, he collaborated with Google as a student researcher and research intern, focusing on the optimization of transformer-based models.

In 2019, he was a research intern at NVIDIA, where his work focused on meta-reinforcement learning. Earlier in his career, in 2018, he served as a research assistant at Microsoft Research in Beijing, contributing to projects in artificial intelligence.

In 2025, Ren became affiliated with Meta’s Superintelligence Labs, a research initiative established to investigate artificial general intelligence. His transition to the lab occurred during a broader expansion effort involving the recruitment of AI researchers from across the industry. ^{[3] [5]}

Achievements

During his time at OpenAI, Hongyu Ren was involved in the co-creation and development of several notable AI models:

• o-series models;
• GPT-4o;
• o1-mini, an efficient model focused on STEM reasoning, math, and coding;
• o3-mini, an efficient reasoning model designed for tasks requiring hard reasoning, web search, and research knowledge;
• o4-mini, a model with text and visual reasoning capabilities He also helped lead post-training efforts for advanced reasoning models at OpenAI. ^{[2] [3]}

Awards

Hongyu Ren has received several fellowships:

• Apple PhD Fellowship;
• Baidu PhD Fellowship;
• Masason Foundation Fellowship. ^{[1] [2]}

Teaching

While at Stanford University, Hongyu Ren served as a teaching assistant for computer science courses.

• Head TA for CS246 Mining Massive Datasets (Winter 2022);
• TA for CS231n Deep Learning for Computer Vision (Spring 2022);
• TA for CS224w Machine Learning on Graphs (Fall 2021). ^[1]