The paper investigates why large language model (LLM) agents fail on long, multi‑step production workflows despite high benchmark success. By testing nine models (1.2 B–671 B parameters) across six task families and multiple horizons, the authors find that task success follows a geometric decay governed by a per‑step reliability that never reaches 1, leading to inevitable collapse for long horizons. The degradation is driven mainly by step count rather than context length, and the study quantifies a significant gap between benchmark and production performance, especially for agentic tool‑use tasks.
By Shubhra Mittal
UniACE is a unified framework that standardizes the evaluation of large language model (LLM) agents by representing each benchmark as an instruction–tool–environment triplet and running models through a shared, task‑agnostic harness in isolated runtimes. It preserves native success criteria, offers an offline mode for dynamic‑resource tasks, and standardizes efficiency metrics, execution records, and failure attribution. Applying UniACE to 7 benchmarks across 24 domains and 15 models revealed significant score shifts, ranking reversals, and sensitivity to evidence representation, highlighting the impact of evaluation configuration on reported agent performance.
By Pengyu Zhu, Lijun Li, Yaxing Lyu, Qianxin Luo, Jingyi Yang, Yi Liu, Tingfeng Hui, Xinyu Yuan, Li Sun, Sen Su, Jing Shao
arXiv:2606. 09863v1 Announce Type: new Abstract: LLM agents can fail silently by asserting task completion when the environment state shows otherwise.
By Laksh Advani
arXiv:2603. 28590v3 Announce Type: replace Abstract: Large language models (LLMs) can generate chains of thought (CoTs) that are not always causally responsible for their final outputs.
By Han Wang, Yifan Sun, Brian Ko, Mann Talati, Jiawen Gong, Zimeng Li, Naicheng Yu, Xucheng Yu, Wei Shen, Vedant Jolly, Huan Zhang
arXiv:2607. 20531v1 Announce Type: new Abstract: Large language model (LLM) agents are increasingly deployed over Model Context Protocol (MCP) servers, yet the benchmarks used to evaluate them score the final answer or a fixed "ground-truth" list of tools, both of which are fragile once the underlying data is live and stateful.
By Jerzy Kami\'nski, Ilya Galyukshev, Artem Kuznetsov, Sergey Chuprin, Kirill Redko, Aidar Shumbalov, Anna Kalyuzhnaya
arXiv:2607. 27283v1 Announce Type: new Abstract: Long-horizon benchmarks often show that agents fail more as tasks become longer.
By Chao Peng, Zhiheng Lyu, Peijie Dong, Hande Dong, Qiang Lin
arXiv:2603. 14465v2 Announce Type: replace Abstract: While Large Language Models (LLMs) have evolved into tool-using agents, they remain brittle in long-horizon interactions.
By Shengda Fan, Xuyan Ye, Yupeng Huo, Zhi-Yuan Chen, Yiju Guo, Shenzhi Yang, Wenkai Yang, Shuqi Ye, Jingwen Chen, Haotian Chen, Xin Cong, Yankai Lin
arXiv:2608. 14380v1 Announce Type: new Abstract: Many real-world tasks require LLM agents to interact with their environments over long execution horizons.
By Yu Zhuang, Kefei Chen, Yitong Duan, Shuxin Zheng, Jian Li, Xu-Yao Zhang
arXiv:2606. 01416v1 Announce Type: new Abstract: Tool-augmented large language model (LLM) agents rely on orchestration layers that coordinate planning, retrieval, tool invocation, validation, memory, and recovery.
By Rahul Suresh Babu, Adarsh Agrawal
arXiv:2607. 10059v1 Announce Type: new Abstract: Agent systems based on large language models (LLMs) are increasingly deployed for autonomous tasks, yet existing evaluations mostly focus on task success rather than whether agents know when to abstain.
By Xun Liu, Yi Evie Zhang, Vira Kasprova, Parisa Rabbani, Pardis Sadat Zahraei, Tianyu Zhang, Ali Ebrahimpour-Boroojeny, Varun Chandrasekaran
The paper introduces a diagnostic framework for long‑horizon security LLM agents that uses checkpoints to distinguish failures occurring before and after a model’s capability is exposed, and applies controlled interventions to pinpoint upstream bottlenecks. The methodology is tested on four task families—delayed reuse of discovered information, reuse of observed state, recovery from failed strategies, and decision making after uncertain outcomes—revealing that many failures happen before the agent observes the state it later needs to reuse. Experiments with Gemini 2.5 Flash and Gemini 3.7 Flash show that targeted protocol‑disambiguation guidance can significantly alter state observation rates and that the primary source of failure can shift across model generations, underscoring the need for fine‑grained failure diagnostics rather than relying solely on overall task success.
By Wei Shao, Chongzhou Fang, Zuxiong Tan, Zequan Liang, Setareh Rafatirad, Avesta Sasan, Houman Homayoun
CivBench is an open‑source benchmark that evaluates language‑model agents in the long‑horizon, tool‑mediated game Civilization VI using the Model Context Protocol (MCP). Each episode lasts over 300 turns, generating thousands of tool calls across a 76‑tool action space, and includes a narration layer that translates visual game state into structured text. The study characterises agent behaviour across four model families, introducing Proactive Monitoring Rate (PMR) and RAG@10 as interface‑level metrics, and finds that agents often under‑monitor strategic state and fail to execute near‑term commitments despite tool access and explicit guidance.
By Austin Tudor David Andrews, Liam Wilkinson, Jamie Heagerty, Harry Coppock, Jakob Nicolaus Foerster, Rui Ponte Costa