The paper introduces "environment evolution," a method that incrementally raises the difficulty of interactive environments off‑policy, scheduling their generation across training generations to supply continuous learning signals. It derives three evolution directions tied to a multi‑turn learning objective and implements them via a loop‑engineered multi‑agent harness. Experiments with models such as Hy4 preview, Claude Opus 5, GPT‑5.6 Sol, Qwen3.6‑27B, and Qwen3.6‑35B‑A3B demonstrate that this approach consistently creates harder environments and boosts terminal‑agent performance on Terminal‑Bench 2.1 by 14.4–18.0 percentage points.
By Zhiyuan Fan, Tinghao Yu, Yuanjun Cai, Jiang Zhou, Jiangtao Guan, Jincheng Liu, Yun Yang, Dingxin Hu, Zhuo Han, Xing Wu, Feng Zhang, Lilin Wang
arXiv:2606. 03108v1 Announce Type: new Abstract: Autonomous LLM training is often framed as recipe search, which leaves the training harness largely static.
By Guhong Chen, Yingcheng Shi, Yongbin Li, Binhua Li, Xander Xu, Hu Wei, Shiwen Ni, Min Yang, Jieping Ye
The paper introduces "environment evolution," a method that incrementally raises the difficulty of interactive environments for terminal agents using off‑policy updates. It derives three evolution directions tied to a multi‑turn learning objective and implements them via a loop‑engineered multi‑agent harness. Experiments with models such as Hy4 preview, Claude Opus 5, GPT‑5.6 Sol, Qwen3.6‑27B, and Qwen3.6‑35B‑A3B demonstrate that this approach consistently generates harder environments and yields significant performance gains on Terminal‑Bench 2.1.
The paper introduces Algorithm Modelization, a paradigm that transforms iterative optimization algorithms into trainable models by treating algorithm execution trajectories as computational experience. It models optimization as an autoregressive process, using a dimension‑independent population‑temporal architecture to capture population interactions and iterative dynamics. The resulting model learns from heterogeneous teachers, improves through optimization feedback, and generalizes across problems and larger dimensions, enabling algorithmic capability to accumulate and evolve via learning.
By Jian Yao, Zhiyao Zhang, Yichen Li, Xingyu Wu, Kay Chen Tan
arXiv:2607. 21971v1 Announce Type: new Abstract: Test-time scaling through iterative self-evolution with environment feedback, as demonstrated by AlphaEvolve, shows remarkable performance gains.
By Shujin Wu, Cheng Qian, Xiusi Chen, Heng Ji
arXiv:2609.22592v1 Announce Type: cross
Abstract: Training agents with reinforcement learning requires a gym, comprising a task, an executable environment in which the task can be attempted, and a ve...
By Aarati Andrea Noronha, Kavya Ravikumar, Carly Xiaoyu Lin
arXiv:2608. 04934v1 Announce Type: cross Abstract: Training LLM agents commonly relies on supervised fine-tuning from expert trajectories or online reinforcement learning over human-specified tasks with handcrafted verifiers.
By Xuanyu Lei, Yiqi Zhu, Chenliang Li, Kaiming Liu, Peng Li, Ming Yan, Jieping Ye, Ya-Qin Zhang, Yang Liu
arXiv:2610.16523v1 Announce Type: new
Abstract: Training LLM agents to interact with users over multiple turns is challenging because the learning signal depends not only on the agent, but also on th...
By Min Cai, Dan Zhang, Gao Xing, Libin Ma, Zhaohui Xu, Xi Ye
AgentFly is an agentic reinforcement learning framework that introduces a unified resource layer to manage heterogeneous environments—such as sandboxes, model services, and external APIs—through a single scheduling engine. The framework is organized into four layers: an agent layer that abstracts agents, tools, and rewards; a rollout layer that composes agent loops and computes rewards; a context layer that organizes rollouts and injects contextual information; and a low‑level resource layer that handles resource management. AgentFly offers prebuilt tools and environments, demonstrates successful agent training across multiple tasks and models, and provides the first controlled cross‑framework throughput comparison against other agentic RL frameworks.
By Renxi Wang, Rifo Ahmad Genadi, Bilal El Bouardi, Yongxin Wang, Fajri Koto, Zhengzhong Liu, Timothy Baldwin, Haonan Li
Agent0 is a fully autonomous framework that enables large language model agents to evolve without external data by using a multi‑step co‑evolution process. It pits a curriculum agent against an executor agent, both derived from the same base LLM, where the curriculum agent creates increasingly challenging tasks and the executor learns to solve them. By integrating external tools into the executor’s workflow, the system creates a self‑reinforcing cycle that continuously generates high‑quality curricula, leading to significant gains in reasoning performance—an 18% improvement on mathematical reasoning and 24% on general reasoning for the Qwen3‑8B‑Base model.
By Peng Xia, Kaide Zeng, Jiaqi Liu, Can Qin, Fang Wu, Yiyang Zhou, Caiming Xiong, Huaxiu Yao
PhantomEnvironments is a framework that trains large language model agents in synthetic, rule‑generated fictional worlds. By creating multi‑turn reinforcement learning environments where agents search templated articles to answer multi‑hop questions, the approach eliminates the need for costly human data or hallucinated LLM‑generated settings. Agents trained in these zero‑cost, purely rule‑based worlds transfer effectively to real‑world multi‑hop search benchmarks, often surpassing models trained on real data, and demonstrate scalable search behavior that grows linearly with question difficulty.
By Anmol Kabra, Swathi Saravana Selvam, Albert Gong, Chao Wan, Christian Belardi, Dongyoung Go, Katie Z. Luo, Kilian Q. Weinberger
Darwin‑Claw is a new framework for large language model agents that enables environment‑adaptive, self‑evolving behavior. It systematically distills episodic experiences into prototype vectors through a hierarchical experience consolidation system, providing context‑sensitive perception patches and explicit optimization signals for skill refinement. The method is deployable with mainstream agent systems and, according to experiments on WildClawBench, delivers significant performance gains while maintaining controllability and stability.
By Mingze Yin, Xiaohan Wang, Dian Li, Longfei Luo, Jiahong Yan, Yiheng Zhu, Chang-Yu Hsieh