Hongyuan Mei

I am building AGI.

Currently, I am working on an interesting project with a few friends.

Previously, I was a Member of Technical Staff at xAI, leading the AI Experts team. My team and I have made core contributions to Grok reasoning models, including Grok 4, Grok 4 Heavy, and Grok 4.1 Fast. My team shipped Grok in Tesla, and drove deep collaboration across xAI, Tesla, SpaceX, and beyond.

Before joining xAI, I was a Senior Research Scientist at Google DeepMind. Before DeepMind, I was a Research Assistant Professor at Toyota Technological Institute at Chicago (TTIC), a research institute affiliated with University of Chicago. At TTIC, I led the River Lab, conducting research on AI reasoning and planning.

I received PhD in Computer Science from Johns Hopkins University, advised by Jason Eisner. We developed the neural Hawkes process, a model widely used in event sequence modeling and prediction.

news [archive]

Apr 17, 2026 MEMO accepted to ICML 2026
Jan 22, 2026 4 papers accepted to ICLR 2026: ∇-Reasoner (latent gradient descent), Executable Counterfactuals (causal reasoning via code), RLVR Generalization (generalizing RLVR), FlowRL (reward distribution matching)
Nov 19, 2025 Grok 4.1 Fast released, with strong agentic capabilities and token efficiency.
Sep 25, 2025 Creativity or Brute Force? accepted to NeurIPS 2025
Jul 9, 2025 Grok 4 and Grok 4 Heavy released, with SOTA on Humanity’s Last Exam