mlsanityData to 14 Oct 2026

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

arXiv 2510.16720 ↗First posted 19 Oct 2025cs.AIcited by 8

On the map: LLM agents and tool use → General LLM agent frameworks and benchmarks

Abstract (arXiv)

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.

Key facts. Cited by 8 papers in the map (data through 14 Oct 2026). Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use. Builds on 230 of its 322 references that are in the map. First posted on arXiv 19 Oct 2025.

Builds on

230 of 322 references.

Closest ancestors · 15

Same direction, newest first

  1. 2025Chain-of-Agents: End-to-End Agent Foundation Models via Multi-Agent Distillation and Agentic RL6 Aug
  2. 2025RCR-Router: Efficient Role-Aware Context Routing for Multi-Agent LLM Systems with Structured Memory6 Aug
  3. 2025Kimi K2: Open Agentic Intelligence28 Jul
  4. 2025Assemble Your Crew: Automatic Multi-agent Communication Topology Design via Autoregressive Graph Generation24 Jul
  5. 2025MasHost Builds It All: Autonomous Multi-Agent System Directed by Reinforcement Learning10 Jun
  6. 2025Reinforce LLM Reasoning through Multi-Agent Reflection10 Jun
  7. 2025Training LLM-Based Agents with Synthetic Self-Reflected Trajectories and Partial Masking26 May
  8. 2025Agentic Knowledgeable Self-awareness4 Apr
  9. 2025Agent models: Internalizing Chain-of-Action Generation into Reasoning models9 Mar
  10. 2025Agent-R: Training Language Model Agents to Reflect via Iterative Self-Training20 Jan
  11. 2025AgentRefine: Enhancing Agent Generalization through Refinement Tuning3 Jan
  12. 2024MALT: Improving Reasoning with Multi-Agent LLM Training2 Dec
3 more
  1. 2024G-Designer: Architecting Multi-agent Communication Topologies via Graph Neural Networks15 Oct
  2. 2024Mixture-of-Agents Enhances Large Language Model Capabilities7 Jun
  3. 2024MAGDi: Structured Distillation of Multi-Agent Interaction Graphs Improves Reasoning in Smaller Language Models2 Feb

Other references · 215

  1. 2022Training language models to follow instructions with human feedbackcited by 15,971
  2. 2017Proximal Policy Optimization Algorithmscited by 15,215
  3. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  4. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  5. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  6. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  7. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  8. 2020Dense Passage Retrieval for Open-Domain Question Answeringcited by 3,824
  9. 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,796
  10. 2023Generative Agents: Interactive Simulacra of Human Behaviorcited by 3,568
  11. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,455
  12. 2020Longformer: The Long-Document Transformercited by 3,413
38 more
  1. 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,810
  2. 2022Locating and Editing Factual Associations in GPTcited by 2,691
  3. 2022Do As I Can, Not As I Say: Grounding Language in Robotic Affordancescited by 2,348
  4. 2023Efficient Streaming Language Models with Attention Sinkscited by 2,186
  5. 2023AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversationcited by 1,988
  6. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,839
  7. 2021WebGPT: Browser-assisted question-answering with human feedbackcited by 1,745
  8. 2020Rethinking Attention with Performerscited by 1,601
  9. 2024From Local to Global: A Graph RAG Approach to Query-Focused Summarizationcited by 1,453
  10. 2025Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learningcited by 1,444
  11. 2023MemGPT: Towards LLMs as Operating Systemscited by 1,389
  12. 2025s1: Simple test-time scalingcited by 1,341
  13. 2021Improving language models by retrieving from trillions of tokenscited by 1,293
  14. 2025Kimi k1.5: Scaling Reinforcement Learning with LLMscited by 1,210
  15. 2017Curiosity-driven Exploration by Self-supervised Predictioncited by 1,194
  16. 2021Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolationcited by 1,103
  17. 2022Measuring and Narrowing the Compositionality Gap in Language Modelscited by 1,075
  18. 2025A-MEM: Agentic Memory for LLM Agentscited by 1,063
  19. 2023Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debatecited by 1,011
  20. 2022Mass-Editing Memory in a Transformercited by 1,001
  21. 2022Atlas: Few-shot Learning with Retrieval Augmented Language Modelscited by 952
  22. 2023Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotationscited by 928
  23. 2023H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Modelscited by 832
  24. 2023CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquingcited by 783
  25. 2025Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memorycited by 779
  26. 2024Self-Rewarding Language Modelscited by 681
  27. 2025SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-trainingcited by 672
  28. 2025BrowseComp: A Simple Yet Challenging Benchmark for Browsing Agentscited by 647
  29. 2025UI-TARS: Pioneering Automated GUI Interaction with Native Agentscited by 610
  30. 2023YaRN: Efficient Context Window Extension of Large Language Modelscited by 601
  31. 2024SeeClick: Harnessing GUI Grounding for Advanced Visual GUI Agentscited by 567
  32. 2023MemoryBank: Enhancing Large Language Models with Long-Term Memorycited by 556
  33. 2025Process Reinforcement through Implicit Rewardscited by 504
  34. 2025LIMO: Less is More for Reasoningcited by 498
  35. 2025Group-in-Group Policy Optimization for LLM Agent Trainingcited by 497
  36. 2024AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agentscited by 468
  37. 2023LongLLMLingua: Accelerating and Enhancing LLMs in Long Context Scenarios via Prompt Compressioncited by 453
  38. 2024Improve Mathematical Reasoning in Language Models by Automated Process Supervisioncited by 441

50 most cited of 215 All 215 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 6.

8 papers.

  1. 2025Memory in the Age of AI Agentscited by 303
  2. 2026Self-Improvements in Modern Agentic Systems: A Surveycited by 19
  3. 2026HiMAC: Hierarchical Macro-Micro Learning for Long-Horizon LLM Agentscited by 9
  4. 2026Aligning Agentic World Models via Knowledgeable Experience Learningcited by 5
  5. 2025DialogGuard: Multi-Agent Psychosocial Safety Evaluation of Sensitive LLM Responsescited by 3
  6. 2026DeferMem: Query-Time Evidence Distillation via Reinforcement Learning for Long-Term Agent Memorycited by 1
  7. 2026Agyn: An Open-Source Platform for AI Agents with Scalable On-Demand Execution, Agent Definition as a Code, and Zero-Trust Accesscited by 0
  8. 2026Piggybacking on Perception: Stealthy Concurrent Audio Prompt Injections against Multimodal LLM Agentscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026The Path Ahead for Agentic AI: Challenges and Opportunities
  2. 2026Rethinking Agentic Reinforcement Learning In Large Language Models
  3. 2025The Landscape of Agentic Reinforcement Learning for LLMs: A Survey
  4. 2026Agentic Artificial Intelligence (AI): Architectures, Taxonomies, and Evaluation of Large Language Model Agents
  5. 2025A Survey on Agentic Multimodal Large Language Models
  6. 2025A Comprehensive Review of AI Agents: Transforming Possibilities in Technology and Beyond
  7. 2025Adaptation of Agentic AI: A Survey of Post-Training, Memory, and Skills
  8. 2025Agent-R1: A Unified and Modular Framework for Agentic Reinforcement Learning