AgentCPM-Explore: Realizing Long-Horizon Deep Exploration for Edge-Scale Agents
ArXiv 2026. First author. Open-source 4B agent model achieving SOTA on GAIA & HLE, surpassing GPT-5 and Claude-4.5-Sonnet.
ToLeaP: Rethinking Development of Tool Learning with Large Language Models
Submitted to ACL 2026. First author. Unified evaluation platform for tool-learning agents.
Reflective Reinforcement Tool Learning
Submitted to ACL 2026. Reflective reinforcement learning for tool learning.
MiniCPM4: Ultra-Efficient LLMs on End Devices
ArXiv 2025. Team contribution (led MCP agent capabilities). 8,300+ GitHub Stars.
minicpm-team
Collaborative Evolving Strategy for Automatic Data-Centric Development
ArXiv 2024. Co-first author. Collaborative evolving strategy for data-centric development. 11,400+ GitHub Stars.
xu-yang
Towards Data-Centric Automatic Research and Development
ArXiv 2024. First author. RD-Agent for automatic R&D. Featured in Microsoft Build 2025 Keynote. 11,400+ GitHub Stars.