EnvFactory: Scaling Tool-Use Agents via Executable Environments Synthesis and Robust RL

Hugging Face Daily Papers Papers
tool-use reinforcement-learning environment-synthesis agentic-rl llm trajectory-synthesis

Summary

EnvFactory automates the creation of executable tool environments and natural multi-turn trajectories for training LLMs with agentic reinforcement learning, achieving superior performance on benchmarks like BFCLv3 and MCP-Atlas with fewer environments than prior work.

Equipping LLMs with tool-use capabilities via Agentic Reinforcement Learning (Agentic RL) is bottlenecked by two challenges: the lack of scalable, robust execution environments and the scarcity of realistic training data that captures implicit human reasoning. Existing approaches depend on costly real-world APIs, hallucination-prone LLM simulators, or synthetic environments that are often single-turn or depend on pre-collected documents. Moreover, synthetic trajectories are frequently over-specified, resembling instruction sequences rather than natural human intents, reducing their effectiveness for RL training. We introduce EnvFactory, a fully automated framework that addresses both challenges. EnvFactory autonomously explores and verifies stateful, executable tool environments from authentic resources, and synthesizes natural multi-turn trajectories through topology-aware sampling and calibrated refinement, producing grounded queries with implicit intents. Using only 85 verified environments across 7 domains, EnvFactory generates 2,575 SFT and RL trajectories. Despite using significantly fewer environments than prior work, which are often 5 times more, EnvFactory achieves superior training efficiency and downstream performance, improving Qwen3-series models by up to +15% on BFCLv3, +8.6% on MCP-Atlas, and +6% on conversational benchmarks including τ^2-Bench and VitaBench. By fully automating both environment construction and trajectory synthesis, EnvFactory provides a scalable, extensible, and robust foundation for Agentic RL.
Original Article
View Cached Full Text

Cached at: 05/20/26, 06:36 AM

Paper page - EnvFactory: Scaling Tool-Use Agents via Executable Environments Synthesis and Robust RL

Source: https://huggingface.co/papers/2605.18703 Authors:

,

,

,

,

,

,

,

,

,

,

,

,

,

Abstract

EnvFactory automates the creation of executable tool environments and natural multi-turn trajectories for training LLMs with agentic reinforcement learning, achieving superior performance with fewer resources.

Equipping LLMs withtool-use capabilitiesviaAgentic Reinforcement Learning(Agentic RL) is bottlenecked by two challenges: the lack of scalable, robustexecution environmentsand the scarcity of realistic training data that captures implicit human reasoning. Existing approaches depend on costly real-world APIs, hallucination-prone LLM simulators, or synthetic environments that are often single-turn or depend on pre-collected documents. Moreover,synthetic trajectoriesare frequently over-specified, resembling instruction sequences rather than natural human intents, reducing their effectiveness for RL training. We introduce EnvFactory, a fully automated framework that addresses both challenges. EnvFactory autonomously explores and verifies stateful, executable tool environments from authentic resources, and synthesizes natural multi-turn trajectories throughtopology-aware samplingandcalibrated refinement, producinggrounded querieswith implicit intents. Using only 85 verified environments across 7 domains, EnvFactory generates 2,575 SFT andRL trajectories. Despite using significantly fewer environments than prior work, which are often 5 times more, EnvFactory achieves superior training efficiency and downstream performance, improvingQwen3-series modelsby up to +15% onBFCLv3, +8.6% onMCP-Atlas, and +6% on conversational benchmarks including τ^2-Bench andVitaBench. By fully automating both environment construction and trajectory synthesis, EnvFactory provides a scalable, extensible, and robust foundation for Agentic RL.

View arXiv pageView PDFGitHub16Add to collection

Get this paper in your agent:

hf papers read 2605\.18703

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper3

#### LARK-Lab/EnvFactory-1.7B Text Generation• 2B• Updatedabout 19 hours ago • 11 #### LARK-Lab/EnvFactory-4B Text Generation• 4B• Updatedabout 19 hours ago #### LARK-Lab/EnvFactory-8B Text Generation• 8B• Updatedabout 2 hours ago

Datasets citing this paper3

#### LARK-Lab/EnvFactory-RL Viewer• Updatedabout 3 hours ago • 3.09k #### LARK-Lab/EnvFactory-SFT-FILTERED Viewer• Updatedabout 3 hours ago • 26.5k #### LARK-Lab/EnvFactory-SFT-ALL Viewer• Updatedabout 3 hours ago • 53.4k

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2605.18703 in a Space README.md to link it from this page.

Collections including this paper1

Similar Articles

AutoEnvScaling: Automating the Data Flywheel with Terminal Agents

Hugging Face Daily Papers

AutoEnvScaling treats RL environment design as a terminal task, letting a proposer agent automatically build training environments from solver rollouts to automate the data flywheel. The method raises the valid-environment rate up to 25x at 75% lower cost, improves Qwen terminal-agent scores by up to 12.2 points, and enables a single model to co-evolve as both solver and generator, a step toward recursive self-improvement.