Hongyu Ren
Hongyu Ren is an AI researcher whose work connects graph learning with language-model reasoning and post-training. His work includes the Open Graph Benchmark and QA-GNN, a question-answering model that combines language models with knowledge graphs. He earned a Stanford PhD, worked at OpenAI on language-model reasoning, and contributed to o1. Since June 2025, he has worked at Meta Superintelligence Labs after leading a post-training group at OpenAI and co-creating GPT-4o and several OpenAI reasoning models.
- Works
- 39
Publication numbers from OpenAlex, as of 15 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.
Timeline
-
2020
An open graph benchmark
In 2020, he co-authors the Open Graph Benchmark, which provides large-scale graph datasets and a unified evaluation protocol for machine-learning research.
-
2021
Question answering with graphs
In 2021, he co-authors QA-GNN, a model that uses language models and knowledge graphs for question answering.
-
2022
Presenting graph reasoning
In 2022, he presents knowledge-graph completion and multi-hop reasoning work at Stanford’s Graph Learning Workshop.
-
October 2024
Stanford PhD and OpenAI work
In October 2024, he holds a PhD from Stanford and works at OpenAI on language-model reasoning and generalization.
-
October 2024
Contributing to o1
In October 2024, he contributes to OpenAI o1, helping advance its ability to solve complex tasks through structured thought processes.
-
Since June 2025
Joining Meta’s superintelligence team
In June 2025, he joins Meta Superintelligence Labs after leading a post-training group at OpenAI and co-creating GPT-4o, 4o-mini, o1-mini, o3-mini and o4-mini.
Sources and credits
Photo: We have no published photo of Hongyu Ren yet. Send us a free-licensed photo.
- Open Graph Benchmark: Datasets for Machine Learning on Graphs
- QA-GNN: Reasoning with Language Models and Knowledge Graphs for Question Answering
- Stanford Graph Learning Workshop 2022 | Computer Science
- Virtual Event: Learning to Reason with LLMs - Event | OpenAI Forum
- Here Is Everyone Mark Zuckerberg Has Hired So Far for Meta’s ‘Superintelligence’ Team
- Meta taps four OpenAI researchers for Superintelligence team | Fortune