mlsanityData to 7 Oct 2026

EnvScaler: Scaling Tool-Interactive Environments for LLM Agent via Programmatic Synthesis

arXiv 2601.05808 ↗First posted 9 Jan 2026cs.CL cs.AI cs.LGcited by 46

On the map: LLM agents and tool use → Tool learning and function calling

Abstract (arXiv)

Large language models (LLMs) are expected to be trained to act as agents in various real-world environments, but this process relies on rich and varied tool-interaction sandboxes. However, access to real systems is often restricted; LLM-simulated environments are prone to hallucinations and inconsistencies; and manually built sandboxes are hard to scale. In this paper, we propose EnvScaler, an automated framework for scalable tool-interaction environments via programmatic synthesis. EnvScaler comprises two components. First, SkelBuilder constructs diverse environment skeletons through topic mining, logic modeling, and quality evaluation. Then, ScenGenerator generates multiple task scenarios and rule-based trajectory validation functions for each environment. With EnvScaler, we synthesize 191 environments and about 7K scenarios, and apply them to Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) for Qwen3 series models. Results on three benchmarks show that EnvScaler significantly improves LLMs' ability to solve tasks in complex environments involving multi-turn, multi-tool interactions. We release our code and data at https://github.com/RUC-NLPIR/EnvScaler.

Key facts. Cited by 46 papers in the map (data through 7 Oct 2026). Placed in Tool learning and function calling, one of 16 directions in LLM agents and tool use. Builds on 16 of its 32 references that are in the map. First posted on arXiv 9 Jan 2026.

Builds on

16 of 32 references.

Closest ancestors · 3

Same direction, newest first

  1. 2025FunReason-MT Technical Report: Advanced Data Synthesis Solution for Real-world Multi-Turn Tool-use28 Oct
  2. 2025Procedural Environment Generation for Tool-Use Agents21 May
  3. 2023API-Bank: A Comprehensive Benchmark for Tool-Augmented LLMs14 Apr

Other references · 13

  1. 2024LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Modelscited by 1,442
  2. 2025WebThinker: Empowering Large Reasoning Models with Deep Research Capabilitycited by 320
  3. 2024ToolSandbox: A Stateful, Conversational, Interactive Evaluation Benchmark for LLM Tool Use Capabilitiescited by 272
  4. 2025REINFORCE++: Stabilizing Critic-Free Policy Optimization with Global Advantage Normalizationcited by 233
  5. 2025APIGen-MT: Agentic Pipeline for Multi-Turn Data Generation via Simulated Agent-Human Interplaycited by 158
  6. 2024WorldCoder, a Model-Based LLM Agent: Building World Models by Writing Code and Interacting with the Environmentcited by 99
  7. 2025UserBench: An Interactive Gym Environment for User-Centric Agentscited by 69
  8. 2025Nemotron-Research-Tool-N1: Exploring Tool-Using Language Models with Reinforced Reasoningcited by 67
  9. 2025Simulating Environments with Reasoning Models for Agent Trainingcited by 52
  10. 2025MUA-RL: Multi-turn User-interacting Agent Reinforcement Learning for agentic tool usecited by 30
  11. 2025PoE-World: Compositional World Modeling with Products of Programmatic Expertscited by 25
  12. 2025AutoForge: Automated Environment Synthesis for Agentic Reinforcement Learningcited by 14
1 more
  1. 2025World Modelling Improves Language Model Agentscited by 12

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 46.

46 papers.

Sort:

By area:

  1. 2026Agent World Model: Infinity Synthetic Environments for Agentic Reinforcement Learningcited by 51
  2. 2026Code as Agent Harnesscited by 40
  3. 2026Agent-World: Scaling Real-World Environment Synthesis for Evolving General Agent Intelligencecited by 35
  4. 2026ScaleEnv: Scaling Environment Synthesis from Scratch for Generalist Interactive Tool-Use Agent Trainingcited by 23
  5. 2026WebWorld: A Large-Scale World Model for Web Agent Trainingcited by 23
  6. 2026Qwen-AgentWorld: Language World Models for General Agentscited by 22
  7. 2026HarnessForge: Joint Harness and Policy Evolution for Adaptive Agent Systemscited by 20
  8. 2026RLAnything: Forge Environment, Policy, and Reward Model in Completely Dynamic RL Systemcited by 19
  9. 2026EnvFactory: Scaling Tool-Use Agents via Executable Environments Synthesis and Robust RLcited by 11
  10. 2026JIT-Agent: Scaling Harness Intelligence via Just-in-Time Harness Evolutioncited by 10
  11. 2026DIVE: Scaling Diversity in Agentic Task Synthesis for Generalizable Tool Usecited by 9
  12. 2026Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Applicationcited by 8
34 more
  1. 2026Recursive Synthesis for Long-Horizon Terminal Taskscited by 8
  2. 2026AgentDoG 1.5: A Lightweight and Scalable Alignment Framework for AI Agent Safety and Securitycited by 7
  3. 2026EnvHarness: Awakening Static Worlds for Agent Learningcited by 7
  4. 2026Skill Self-Play: Pushing the Frontier of LLM Capability with Co-Evolving Skillscited by 6
  5. 2026LOGIGEN: Logic-Driven Generation of Verifiable Agentic Taskscited by 5
  6. 2026Self-Improvement of Large Language Models: A Technical Overview and Future Outlookcited by 5
  7. 2026Learning to Build the Environment: Self-Evolving Reasoning RL via Verifiable Environment Synthesiscited by 5
  8. 2026AgentSkiller: Scaling Generalist Agent Intelligence through Semantically Integrated Cross-Domain Data Synthesiscited by 3
  9. 2026Bridging the Agent-World Gap: Text World Models for LLM-based Agentscited by 3
  10. 2026TRACE: Capability-Targeted Agentic Trainingcited by 2
  11. 2026Agent-ValueBench: A Comprehensive Benchmark for Evaluating Agent Valuescited by 2
  12. 2026Beyond Simply Environment Scaling: Designing Effective Environment Distributions for Multimodal Agent Learningcited by 2
  13. 2026EnvACE: Internalizing Environment Dynamics via World Rehearsal for Agentic Reinforcement Learningcited by 2
  14. 2026What Makes Good Agentic Data? An ACE Lens on Data Generation for LLM Agentscited by 2
  15. 2026Democratizing Tool Learning with Environments Fully Simulated by a Free 8B Language Modelcited by 1
  16. 2026Scalable Environments Drive Generalizable Agentscited by 1
  17. 2026OPD-Evolver: Cultivating Holistic Agent Evolver via On-Policy Distillationcited by 1
  18. 2026The Path to Self-Evolving Clinical Systems: Scaling Medical Agents from Assistance to Autonomycited by 1
  19. 2026Masked Diffusion Language Models are Strong and Steerable Text-Based World Models for Agentic RLcited by 1
  20. 2026EnvCraft: Synthesizing Executable Environments in Agentic RL for Claw-like Agentcited by 1
  21. 2026EnvSimBench: A Benchmark for Evaluating and Improving LLM-Based Environment Simulationcited by 0
  22. 2026CurateEvo: Data-Curation Evolving for Agentic Post-Trainingcited by 0
  23. 2026ABot-AgentOS: A General Robotic Agent OS with Lifelong Multi-modal Memorycited by 0
  24. 2026AgentOmnia: Scaling Agentic Models for Full-Scenario Applicationscited by 0
  25. 2026AgentMercury: Your Agent Can Synthesize Verifiable Environments for Business Scenarios at scalecited by 0
  26. 2026FinCUABuild: Can Agents Build Reliable Benchmarks for Dynamic Financial Computer Use?cited by 0
  27. 2026Environments as Scaffold: Enriching Feedback to Bootstrap Self-Evolving Agents in Long-Horizon Taskscited by 0
  28. 2026AutoGym: Blueprint-First Generation of Verifiable Agent Gymscited by 0
  29. 2026World State Generatorcited by 0
  30. 2026Verifiable Hidden Dynamics Play: Generating Agentic RL Environments from Solved Mechanismscited by 0
  31. 2026CompoWorld: Compositional Environment Scaling for General Agentscited by 0
  32. 2026WEFT: Scaling Tool-Use Post-Training for General-Purpose Agentscited by 0
  33. 2026SkillGym: Training Skill-Use Agents with Automatic Verifiable Environment Generationcited by 0
  34. 2026AdviSD: Learning to Advise Frontier LLMs via Targeted Multi-Turn Self-Distillationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026OpaqueToolsBench: Learning Nuances of Tool Behavior Through Interaction
  2. 2025Feedback-Driven Tool-Use Improvements in Large Language Models via Automated Build Environments
  3. 2025Generalizable End-to-End Tool-Use RL with Synthetic CodeGym
  4. 2026Teaching LLMs to Learn Tool Trialing and Execution through Environment Interaction
  5. 2026Scaling Agentic Capabilities, Not Context: Efficient Reinforcement Finetuning for Large Toolspaces
  6. 2024A Review of Prominent Paradigms for LLM-Based Agents: Tool Use (Including RAG), Planning, and Feedback Learning
  7. 2024Learning to Use Tools via Cooperative and Interactive Agents
  8. 2026UniToolCall: Unifying Tool-Use Representation, Data, and Evaluation for LLM Agents