AgentPerturb introduces a systematic framework to evaluate large language model agents under controlled interaction perturbations, addressing the lack of unified evaluation methods for noisy real‑world tasks. The study quantifies robustness across frontier models, identifies task‑specific vulnerabilities, and shows that agents often fail to detect altered inputs, propagating errors and incurring higher computational costs. The authors provide a benchmark‑construction pipeline and plan to release code and data to support future robustness‑oriented training.
By Yuanjian Xu, Jinyi Han, Jianing Hao, Yuxuan Sun, Fanyang Lu, Richeng Xuan, Guang Zhang, Zhichao Hu
The paper introduces STEPGATE, an uncertainty‑aware handoff framework that evaluates each step of a small language model (SLM) agent and selectively escalates difficult steps to a stronger model. On a 52‑task single‑step benchmark, the Qwen2.5‑1.5B/7B pair achieved 82.7% task success with only 30.8% escalation, outperforming local‑only and random escalation baselines. In multi‑turn tests, STEPGATE reached 69.0% trajectory success and 84.0% action success while using only 30.0% cloud actions, demonstrating that step‑level escalation can close much of the performance gap to a stronger backend with fewer remote tokens.
By Abolfazl Younesi
arXiv:2608. 03222v1 Announce Type: cross Abstract: Software engineering (SWE) agents resolve repository-level issues through long trajectories that grow increasingly expensive as context accumulates.
By Chenyu Wang, Yunbo Lyu, Junda He, Zhou Yang, Chenxing Zhong, Yaniv Harel, David Lo
Software engineering (SWE) agents resolve repository-level issues through long trajectories that grow increasingly expensive as context accumulates. Failed runs tend to be longer and exhibit redundant exploration or looping, suggesting that some failures may be detectable before completion.
arXiv:2610.09115v1 Announce Type: cross
Abstract: Steering an LLM agent means deciding whether to correct it, at which step, and with which mechanism. Uncertainty is often used to decide when to corr...
By Hanwen Li, Jinhao Duan, Guanhua Zhu, Junchi Lu, Bo Shen, Chenxi Yuan, Kaidi Xu
arXiv:2608.29685v1 Announce Type: new
Abstract: Early failure prediction is important for long-horizon agents, as it enables timely intervention and can reduce inference and tool-use costs. Uncertain...
By Zongyue Li, Chengyue Yu, Lei Zang, Chenyi Zhuang, Linjian Mo, Leilei Gan
The paper proposes CANOPY, a minimalist reinforcement learning protocol that addresses two common pitfalls—signal starvation and policy drift—in outcome‑only RL for long‑horizon interactive tasks. By scaling same‑task exploration, keeping updates on‑policy, and anchoring updates with KL divergence, CANOPY enables a Qwen3‑14B agent to achieve top leaderboard results on the AppWorld coding benchmark without auxiliary supervision or elaborate scaffolding. The approach also improves performance on SWE‑bench for a Qwen3.5‑9B model.
By Liming Pu, Xiaoxia Li, Yifu Liu, Teng Cao, Bin Yang
arXiv:2602. 11619v2 Announce Type: replace Abstract: Running the same LLM agent on identical inputs yields 2.
By Aman Mehta
arXiv:2608.05446v2 Announce Type: replace
Abstract: Long-horizon LLM agents increasingly rely on external execution support to maintain state, track progress, recover from failures, and reuse experie...
By Xuying Ning, Dongqi Fu, Tianxin Wei, Yuanchen Bei, Xiyuan Yang, Wujiang Xu, Yueqi Song, Bingxuan Li, Zihao Li, Hanqing Zeng, Xiang Shen, Yajuan Wang, Yifan Wu, Qifan Wang, Jiayi Liu, Hong Li, Yinglong Xia, Xiangjun Fan, Hanghang Tong, Jingrui He
Mingbird is a local‑first agent harness designed for small open‑weight language models (2–9 B) that run on ordinary laptops. It introduces ten mechanisms—such as a byte‑level net‑zero prefill budget, a finish gate that re‑reads the task before accepting completion, and signature‑level loop detection—to address common failure modes that arise from the harness rather than the model itself. In controlled experiments on the LRAB benchmark and the $ au^2$‑bench, Mingbird achieves higher overall scores (0.886 and 0.856 respectively) compared to other harnesses, and its ablation studies show that each mechanism contributes measurable performance gains.
By Hao Wang, Ting Huang
arXiv:2603.01209v3 Announce Type: replace
Abstract: In CodeAct, language-model agents write Python that calls tools and use execution feedback to choose actions. Persistent runtimes preserve Python v...
By Victor May, Van Khue Nguyen, Aaditya Salgarkar, Yishan Wang, Diganta Misra, Huu Nguyen
The paper introduces Traverse, a benchmark of 2,518 agent trajectories and 6,967 annotated mistakes across software engineering, computer use, and science tasks, revealing that failures often go unrecovered and can cause irreversible harm before a run is deemed successful. It shows that human judges struggle to detect the first mistake in most runs, while a 4‑billion‑parameter verifier called Scout can locate failures more effectively and improve task success when used to select among candidate runs. The study demonstrates that making failure detection inexpensive and reliable can enable long‑horizon agents to learn from their own mistakes and increase trustworthiness in autonomous AI.
By Salman Rahman, Yubin Kim, Mihir Parmar, A. Ali Heydari, Genglin Liu, Simon A. Lee, Weizhi Zhang, Arian Hosseini, Ahmed A. Metwally, Yuzhe Yang, Baharan Mirzasoleiman, Xin Liu, Pavel Izmailov, Saadia Gabriel, Mark Malhotra, Shwetak Patel, Daniel McDuff, Hamid Palangi