| Feb 09, 2026 | 🚀 New! We’ve released OpenResearcher — a fully open, state-of-the-art deep research agent (30B-A3B). Model, data, code — everything is open. Try demo now! |
| Jan 27, 2026 | Two papers accepted at ICLR 2026: AgentFlow (Oral) on in-the-flow agentic system optimization, and ImagenWorld on stress-testing image generation models! |
| Dec 13, 2025 | 🏆 Excited that AgentFlow won Best Paper Runner-up at the Effective Reasoning Workshop! |
| Dec 12, 2025 | 🔥 Thrilled to share that StructEval has been acceptecd to TMLR! |
| Oct 19, 2025 | Thrilled to announce that AgentFlow has been accepted to NeurIPS 2025 Workshop Oral (Top-3) and ranked #2 Paper of the Day on Hugging Face 🤗! |
| Oct 05, 2025 | Thrilled to introduce AgentFlow — a trainable, tool-integrated agentic framework that directly optimizes agents within the system in an online fashion using Flow-GRPO 🌀💫, achieving superior tool use 🛠 and long-horizon reasoning 🧠 across diverse domains! |
| Oct 02, 2025 | Excited to share VideoScore2 — bringing RL to generative video evaluation with scoring and rich reasoning traces! |
| Aug 22, 2025 | 🎉 VerlTool is approaching 550+ stars on GitHub! It’s becoming a key framework for Tool-Agent RL training. |
| Aug 18, 2025 | VerlTool’s tech report (co-first author) is out! Please see on Hugging Face Daily Paper! |
| Aug 15, 2025 | 🚀 Starting my CS PhD Journey at Teaxs A&M University! |
| Aug 13, 2025 | 🤝 Our paper GReF (co-first author) has been accepted to CIKM 2026. This recommendation reranking LLM project was conducted during my internship at Kuaishou. |
| Jun 11, 2025 | 🏛️ I start my exciting journey as a research intern at Stanford University, working with Prof. James Zou, Prof. Yejin Choi. |