# OpenForgeRL: Train Harness-native Agents in Any Environment

arXiv: https://arxiv.org/abs/2607.21557 · HTML: https://mlsanity.com/paper/2607.21557/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 30 of its 53 references that are in the map.
- First posted on arXiv 23 Jul 2026.

## Builds on (most cited first, 30 of 30)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [Voyager: An Open-Ended Embodied Agent with Large Language Models](https://mlsanity.com/paper/2305.16291/) (2023)
- [HybridFlow: A Flexible and Efficient RLHF Framework](https://mlsanity.com/paper/2409.19256/) (2024)
- [AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation](https://mlsanity.com/paper/2308.08155/) (2023)
- [SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering](https://mlsanity.com/paper/2405.15793/) (2024)
- [Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning](https://mlsanity.com/paper/2503.09516/) (2025)
- [Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Models](https://mlsanity.com/paper/2506.05176/) (2025)
- [CogAgent: A Visual Language Model for GUI Agents](https://mlsanity.com/paper/2312.08914/) (2023)
- [Executable Code Actions Elicit Better LLM Agents](https://mlsanity.com/paper/2402.01030/) (2024)
- [Group-in-Group Policy Optimization for LLM Agent Training](https://mlsanity.com/paper/2505.10978/) (2025)
- [WebVoyager: Building an End-to-End Web Agent with Large Multimodal Models](https://mlsanity.com/paper/2401.13919/) (2024)
- [AReaL: A Large-Scale Asynchronous Reinforcement Learning System for Language Reasoning](https://mlsanity.com/paper/2505.24298/) (2025)
- [Agent S: An Open Agentic Framework that Uses Computers Like a Human](https://mlsanity.com/paper/2410.08164/) (2024)
- [An Illusion of Progress? Assessing the Current State of Web Agents](https://mlsanity.com/paper/2504.01382/) (2025)
- [AgentTrek: Agent Trajectory Synthesis via Guiding Replay with Web Tutorials](https://mlsanity.com/paper/2412.09605/) (2024)
- [OpenClaw-RL: Train Any Agent Simply by Talking](https://mlsanity.com/paper/2603.10165/) (2026)
- [Claw-Eval: Towards Trustworthy Evaluation of Autonomous Agents](https://mlsanity.com/paper/2604.06132/) (2026)
- [Proposer-Agent-Evaluator(PAE): Autonomous Skill Discovery For Foundation Model Internet Agents](https://mlsanity.com/paper/2412.13194/) (2024)
- [AgentSynth: Scalable Task Generation for Generalist Computer-Use Agents](https://mlsanity.com/paper/2506.14205/) (2025)
- [Fara-7B: An Efficient Agentic Model for Computer Use](https://mlsanity.com/paper/2511.19663/) (2025)
- [SWE-rebench V2: Language-Agnostic SWE Task Collection at Scale](https://mlsanity.com/paper/2602.23866/) (2026)
- [WebGym: Scaling Training Environments for Visual Web Agents with Realistic Tasks](https://mlsanity.com/paper/2601.02439/) (2026)
- [Polar: Agentic RL on Any Harness at Scale](https://mlsanity.com/paper/2605.24220/) (2026)
- [Dyna-Mind: Learning to Simulate from Experience for Better AI Agents](https://mlsanity.com/paper/2510.09577/) (2025)
- [Orchard: An Open-Source Agentic Modeling Framework](https://mlsanity.com/paper/2605.15040/) (2026)
- [OpenWebRL: Demystifying Online Multi-turn Reinforcement Learning for Visual Web Agents](https://mlsanity.com/paper/2606.02031/) (2026)
- [Synthetic Computers at Scale for Long-Horizon Productivity Simulation](https://mlsanity.com/paper/2604.28181/) (2026)

## Built on by (most cited first, 7 of 7)

- [LEGO-RL: Harness-Native Reinforcement Learning for Coding Agents](https://mlsanity.com/paper/2608.17393/) (2026)
- [Ecdysis: Efficient and Effective Training of Runtime Harnesses for LLM Agents](https://mlsanity.com/paper/2609.11677/) (2026)
- [ClawGym II: Exploring Black-Box RL on Agent Harness](https://mlsanity.com/paper/2608.16798/) (2026)
- [Multi-Harness Reinforcement Learning for Coding Agents: Training Gains and Harness Dependence](https://mlsanity.com/paper/2609.04518/) (2026)
- [VISTA: A Visual Harness for Reasoning in an Interactive World](https://mlsanity.com/paper/2610.02200/) (2026)
- [SCLATE: a Substrate for Continual-Learning Agent Training and Evaluation](https://mlsanity.com/paper/2609.32391/) (2026)
- [SCOUT: Synergizing Reasoning and Tool-Use for Computer-Use Safety](https://mlsanity.com/paper/2609.36201/) (2026)

## Abstract (arXiv)

Modern AI agents rely on elaborate inference harnesses such as Claude Code, Codex, and OpenClaw to drive multi-turn reasoning, tool use, and access to external systems. While powerful, these complex harnesses also make agents hard to train end-to-end with open infrastructure, whose SFT/RL stacks cannot natively express stateful, multi-process harness inference. To address this, we present OpenForgeRL, an open-source framework for training harness-based agents end-to-end in diverse environments. OpenForgeRL achieves this with a lightweight proxy that serves the harness's model calls while recording them as training data for a standard RL codebase (e.g., veRL), and a Kubernetes orchestrator that runs each rollout in its own remote container, together enabling training on any harness in any environment at scale. By decoupling training and inference, OpenForgeRL allows researchers to easily train, study, and improve agents directly in the real harnesses and environments they are deployed with. We validate our framework across diverse, complex harnesses and environments, spanning tool/claw-based agents and multimodal GUI browser- and computer-use agents. Using only hundreds to a few thousand tasks, OpenForgeClaw reaches 31.7 pass^3 and 55.9 pass@3 on ClawEval and 33.7 on QwenClawBench. OpenForgeGUI reaches 37.7 on OSWorld-Verified, 63.0 on Online-Mind2Web, and 72.3 on WebVoyager. Both outperform open baselines of similar size on nearly all benchmarks, and in the GUI setting match or surpass models several times larger. Beyond benchmarks, we analyze how harness choice (e.g., ZeroClaw, OpenClaw, Codex) and RL shape agent behavior. We find that some harnesses are substantially harder to learn than others, and that RL improves agentic reliability, such as self-verification, tool coverage, and completing multi-step plans, though critical abilities such as error recovery remain weak.