mlsanityData to 7 Oct 2026

Let It Flow: Agentic Crafting on Rock and Roll, Building the ROME Model within an Open Agentic Learning Ecosystem

arXiv 2512.24873 ↗First posted 31 Dec 2025cs.AI cs.CLcited by 28

On the map: LLM agents and tool use → Reinforcement learning for LLM agents

Abstract (arXiv)

Agentic crafting requires LLMs to operate in real-world environments over multiple turns by taking actions, observing outcomes, and iteratively refining artifacts. Despite its importance, the open-source community lacks a principled, end-to-end ecosystem to streamline agent development. We introduce the Agentic Learning Ecosystem (ALE), a foundational infrastructure that optimizes the production pipeline for agentic model. ALE consists of three components: ROLL, a post-training framework for weight optimization; ROCK, a sandbox environment manager for trajectory generation; and iFlow CLI, an agent framework for efficient context engineering. We release ROME, an open-source agent grounded by ALE and trained on over one million trajectories. Our approach includes data composition protocols for synthesizing complex behaviors and a novel policy optimization algorithm, Interaction-Perceptive Agentic Policy Optimization (IPA), which assigns credit over semantic interaction chunks rather than individual tokens to improve long-horizon training stability. Empirically, we evaluate ROME within a structured setting and introduce Terminal Bench Pro, a benchmark with improved scale and contamination control. ROME demonstrates strong performance across benchmarks like SWE-bench Verified and Terminal Bench, proving the effectiveness of ALE.

Key facts. Cited by 28 papers in the map (data through 7 Oct 2026). Placed in Reinforcement learning for LLM agents, one of 16 directions in LLM agents and tool use. Builds on 38 of its 57 references that are in the map. First posted on arXiv 31 Dec 2025.

Builds on

38 of 57 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025RollArt: Disaggregated Multi-Task Agentic RL Training at Scale27 Dec
  2. 2025Klear-AgentForge: Forging Agentic Intelligence through Posttraining Scaling8 Nov

Other references · 36

  1. 2017Proximal Policy Optimization Algorithmscited by 14,441
  2. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,767
  3. 2023SWE-bench: Can Language Models Resolve Real-World GitHub Issues?cited by 3,193
  4. 2023Retrieval-Augmented Generation for Large Language Models: A Surveycited by 2,185
  5. 2019Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelismcited by 1,829
  6. 2023SGLang: Efficient Execution of Structured Language Model Programscited by 1,107
  7. 2023GAIA: a benchmark for General AI Assistantscited by 1,084
  8. 2024OpenHands: An Open Platform for AI Software Developers as Generalist Agentscited by 956
  9. 2025Group Sequence Policy Optimizationcited by 788
  10. 2025AlphaEvolve: A coding agent for scientific and algorithmic discoverycited by 763
  11. 2024A Survey on Large Language Models for Code Generationcited by 646
  12. 2025τ2-Bench: Evaluating Conversational Agents in a Dual-Control Environmentcited by 558
24 more
  1. 2025Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoningcited by 540
  2. 2024Agentless: Demystifying LLM-based Software Engineering Agentscited by 400
  3. 2023Large Language Models for Software Engineering: A Systematic Literature Reviewcited by 389
  4. 2025VAPO: Efficient and Reliable Reinforcement Learning for Advanced Reasoning Taskscited by 281
  5. 2016Safe and Efficient Off-Policy Reinforcement Learningcited by 260
  6. 2025Model Context Protocol (MCP): Landscape, Security Threats, and Future Research Directionscited by 257
  7. 2025SWE-RL: Advancing LLM Reasoning via Reinforcement Learning on Open Software Evolutioncited by 246
  8. 2017A Survey of Machine Learning for Big Code and Naturalnesscited by 220
  9. 2025BrowseComp-ZH: Benchmarking Web Browsing Ability of Large Language Models in Chinesecited by 139
  10. 2025Mobile-Agent-E: Self-Evolving Mobile Assistant for Complex Taskscited by 128
  11. 2025Geometric-Mean Policy Optimizationcited by 107
  12. 2025Reinforcement Learning with Action Chunkingcited by 97
  13. 2025Reinforcement Learning Optimization for Large-Scale Learning: An Efficient and User-Friendly Scaling Librarycited by 94
  14. 2025Soft Adaptive Policy Optimizationcited by 89
  15. 2025Tapered Off-Policy REINFORCE: Stable and efficient reinforcement learning for LLMscited by 61
  16. 2025Seed-Coder: Let the Code Model Curate Data for Itselfcited by 46
  17. 2020Discount Factor as a Regularizer in Reinforcement Learningcited by 44
  18. 2025Stabilizing Reinforcement Learning with LLMs: Formulation and Practicescited by 38
  19. 2025History Rhymes: Accelerating LLM Reinforcement Learning with RhymeRLcited by 35
  20. 2025Every Activation Boosted: Scaling General Reasoner to 1 Trillion Open Language Foundationcited by 32
  21. 2025RollPacker: Mitigating Long-Tail Rollouts for Fast, Synchronous RL Post-Trainingcited by 31
  22. 2024MTU-Bench: A Multi-granularity Tool-Use Benchmark for Large Language Modelscited by 17
  23. 2025DeepTravel: An End-to-End Agentic Reinforcement Learning Framework for Autonomous Travel Planning Agentscited by 17
  24. 2024Analyzing and Bridging the Gap between Maximizing Total Reward and Discounted Reward in Deep Reinforcement Learningcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 28.

28 papers.

Sort:

By area:

  1. 2026SWE-CI: Evaluating Agent Capabilities in Maintaining Codebases via Continuous Integrationcited by 18
  2. 2026SkillsVote: Lifecycle Governance of Agent Skills from Collection, Recommendation to Evolutioncited by 18
  3. 2026ThunderAgent: A Simple, Fast and Program-Aware Agentic Inference Systemcited by 17
  4. 2026AgentHazard: A Benchmark for Evaluating Harmful Behavior in Computer-Use Agentscited by 16
  5. 2026Terminal Wrench: A Dataset of 331 Reward-Hackable Environments and 3,632 Exploit Trajectoriescited by 14
  6. 2026Crab: A Semantics-Aware Checkpoint/Restore Runtime for Agent Sandboxescited by 11
  7. 2026Exact Is Easier: Credit Assignment for Cooperative LLM Agentscited by 9
  8. 2026RTMC: Step-Level Credit Assignment via Rollout Treescited by 9
  9. 2026Hardening Agent Benchmarks with Adversarial Hacker-Fixer Loopscited by 8
  10. 2026The Scaffold Effect in Coding Agents: Harness Choice as a Hidden Variable in Coding-Agent Evaluationcited by 8
  11. 2026Orchard: An Open-Source Agentic Modeling Frameworkcited by 7
  12. 2026LiteCoder-Terminal: Scaling Long-Horizon Terminal Environments for Learning Language Agentscited by 7
16 more
  1. 2026The Meta-Agent Challenge: Are Current Agents Capable of Autonomous Agent Development?cited by 7
  2. 2026Missing Old Logits in Asynchronous Agentic RL: Semantic Mismatch and Repair Methods for Off-Policy Correctioncited by 4
  3. 2026BraveGuard: From Open-World Threats to Safer Computer-Use Agentscited by 4
  4. 2026What Makes a Good Terminal-Agent Benchmark Task: A Guideline for Adversarial, Difficult, and Legible Evaluation Designcited by 3
  5. 2026Learning Agentic Policy from Action Guidancecited by 3
  6. 2026How Does Reasoning Flow? Tracing Attention-Induced Information Flow for Targeted RL in LLMscited by 3
  7. 2026LEGO-RL: Harness-Native Reinforcement Learning for Coding Agentscited by 3
  8. 2026Learning Generalizable Behaviors for Terminal Agentscited by 1
  9. 2026SkillGLoW: Procedural-Family Skill Consolidation for Self-Improving Agents on Long-Horizon Task Streamscited by 1
  10. 2026T1: Terminal Agent Reinforcement Learning for Long-Horizon Taskscited by 1
  11. 2026Detecting Intrinsic and Instrumental Self-Preservation in Autonomous Agents: The Unified Continuation-Interest Protocolcited by 0
  12. 2026ARL-Tangram: Unleash the Resource Efficiency in Agentic Reinforcement Learningcited by 0
  13. 2026Rollout Pass-Rate Control: Steering Binary-Reward RL Toward Its Most Informative Regimecited by 0
  14. 2026AgentOmnia: Scaling Agentic Models for Full-Scenario Applicationscited by 0
  15. 2026One to More, More to One: Category-Aware Iterative Expert Training for Software Engineering Agentscited by 0
  16. 2026Comprehension Audits to Mitigate Risks from Automated AI Researchcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025CrafterDojo: A Suite of Foundation Models for Building Open-Ended Embodied Agents in Crafter
  2. 2025OpenHA: A Series of Open-Source Hierarchical Agentic Models in Minecraft
  3. 2022MineDojo: Building Open-Ended Embodied Agents with Internet-Scale Knowledge
  4. 2026Agent Skill Evaluation and Evolution: Frameworks and Benchmarks
  5. 2026Mock Worlds, Real Skills: Building Small Agentic Language Models with Synthetic Tasks, Simulated Environments, and Rubric-Based Rewards
  6. 2025ToolBrain: A Flexible Reinforcement Learning Framework for Agentic Tools
  7. 2025Agent-R1: A Unified and Modular Framework for Agentic Reinforcement Learning
  8. 2026OpenGame: Open Agentic Coding for Games