Shypd.aiThe largest open index of AI toolsAgentCPM is an end-to-end infrastructure for training and evaluating various LLM agents. It includes AgentCPM-Explore, a 4B parameter deep-search agent, and AgentCPM-Report, an 8B parameter deep-research agent for report generation.
Trending
About
AgentCPM is an open-source infrastructure developed by THUNLP, Renmin University of China, ModelBest, and the OpenBMB community, designed for training and evaluating various LLM agents. It addresses challenges in real-world applications such as limited long-horizon capability, autonomy, and generalization. The platform features AgentCPM-Explore, a 4B parameter deep-search LLM agent that achieves state-of-the-art performance on long-horizon benchmarks, and AgentCPM-Report, an 8B parameter deep-research LLM agent built on MiniCPM4.1-8B, capable of generating comprehensive reports comparable to top commercial systems. AgentCPM provides end-to-end open-source code for training, inference, and evaluation, along with a unified tool sandbox environment (AgentDock) for collaborative multi-model and multi-tool setups.
Capabilities
Pricing & Plans
Open Source
Free
FAQs
Trending
