超越流水線：邁向模型原生智能代理的範式轉變綜述

Beyond Pipelines: A Survey of the Paradigm Shift toward Model-Native Agentic AI

October 19, 2025

作者: Jitao Sang, Jinlin Xiao, Jiarun Han, Jilin Chen, Xiaoyi Chen, Shuyu Wei, Yongjie Sun, Yuhang Wang

cs.AI

摘要

智能代理AI的快速發展標誌著人工智慧進入了一個新階段，在這個階段，大型語言模型（LLMs）不再僅僅是回應，而是能夠行動、推理和適應。本調查追溯了構建智能代理AI的範式轉變：從基於管線的系統——其中規劃、工具使用和記憶由外部邏輯協調——到新興的模型原生範式，這些能力被內化於模型的參數之中。我們首先將強化學習（RL）定位為推動這一範式轉變的算法引擎。通過將學習從模仿靜態數據重新定義為結果驅動的探索，RL支撐了跨語言、視覺和具身領域的LLM + RL + 任務的統一解決方案。基於此，本調查系統性地回顧了每項能力——規劃、工具使用和記憶——如何從外部編寫的模塊演變為端到端學習的行為。此外，它探討了這一範式轉變如何重塑了主要的代理應用，特別是強調長遠推理的深度研究代理和強調具身交互的GUI代理。最後，我們討論了智能代理能力的持續內化，如多代理協作和反思，以及系統層和模型層在未來智能代理AI中角色的演變。這些發展共同勾勒出模型原生智能代理AI作為一個整合學習與交互框架的連貫軌跡，標誌著從構建應用智能的系統向開發通過經驗增長智能的模型的轉變。

English

The rapid evolution of agentic AI marks a new phase in artificial intelligence, where Large Language Models (LLMs) no longer merely respond but act, reason, and adapt. This survey traces the paradigm shift in building agentic AI: from Pipeline-based systems, where planning, tool use, and memory are orchestrated by external logic, to the emerging Model-native paradigm, where these capabilities are internalized within the model's parameters. We first position Reinforcement Learning (RL) as the algorithmic engine enabling this paradigm shift. By reframing learning from imitating static data to outcome-driven exploration, RL underpins a unified solution of LLM + RL + Task across language, vision and embodied domains. Building on this, the survey systematically reviews how each capability -- Planning, Tool use, and Memory -- has evolved from externally scripted modules to end-to-end learned behaviors. Furthermore, it examines how this paradigm shift has reshaped major agent applications, specifically the Deep Research agent emphasizing long-horizon reasoning and the GUI agent emphasizing embodied interaction. We conclude by discussing the continued internalization of agentic capabilities like Multi-agent collaboration and Reflection, alongside the evolving roles of the system and model layers in future agentic AI. Together, these developments outline a coherent trajectory toward model-native agentic AI as an integrated learning and interaction framework, marking the transition from constructing systems that apply intelligence to developing models that grow intelligence through experience.