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
Enterprise practitioners read agent leaderboards as if they ranked agent capability. We show, across three open agent-trace benchmarks (TheAgentCompany, $τ^2$-bench, and AppWorld), that the agent main effect accounts for less than 3% of total variance in every dataset and check type, while the agent-by-task interaction accounts for 7-23%.
arXiv:2608. 11323v1 Announce Type: new Abstract: Enterprise practitioners read agent leaderboards as if they ranked agent capability.
By Vasundra Srinivasan
arXiv:2606. 29654v1 Announce Type: new Abstract: Multi-agent deliberation among LLMs can improve reasoning, but deployment requires deciding when the current answer is reliable enough to act on and when it should be escalated to human review.
By Mengdie Flora Wang, Haochen Xie, Guanghui Wang, Devin Zhang, Jae Oh Woo
arXiv:2606. 01365v1 Announce Type: new Abstract: Tool-using multi-agent large language model (LLM) systems spend computation through model tokens, tool calls, retries, and code execution before producing an answer.
By Xianyou Li, Weiran Yan, Yichao Wu, Penghao Liang, Mengwei Yuan, Jianan Liu, Jing Yang
AgentAudit is an open, extensible framework that evaluates the full lifecycle of AI agents, assessing planning, tool selection, execution, memory, and reasoning across ten dimensions such as instruction integrity, security, and alignment. Unlike existing benchmarks that focus on single aspects, AgentAudit analyzes the entire execution trace to attribute failures to specific stages. The framework was tested on five large language models, revealing significant differences in trustworthiness even among models with similar task‑completion performance.
By Shrey Nag, Sachita, Abhishek Kumar Singh, Lipi Goel, Rajeshwar Singh Janwar
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
arXiv:2608. 15286v1 Announce Type: cross Abstract: We introduce AgentRelBench, an environment-agnostic reliability instrument that computes ground-truth, severity-priced damage from database state diffs across repeated runs, with no LLM in the measurement path, demonstrated on EnterpriseOps-Gym.
By Shiven Khurdi
arXiv:2607. 02436v1 Announce Type: cross Abstract: Agentic coding assistants are increasingly given extra capabilities, such as browser based testing tools and design oriented system prompts, on the assumption that more capability yields better software.
By Achint Mehta
The paper investigates the reliability of tool‑using agents, focusing on two failure modes: selecting the wrong tool and constructing incorrect arguments. It introduces a correct‑invocation rate metric to distinguish these errors and evaluates five open‑weight models on multi‑step tasks up to depth 8, finding that by depth 6 about 70% of a model’s clean‑context capability is lost due to earlier mistakes. The study reveals that exact‑match scoring against a fixed gold trajectory forces severity and recovery parameters to extreme values, and proposes a conditional‑on‑state scoring remedy that yields more realistic severity estimates.
By Afiya Noorain, Subhranshu Mohanty, Amritesh Banerjee, Abhijit Dasgupta
TRACE (TRAjectory-Contrastive Evolution) is a self‑evolving skill bank that improves the consistency and limit‑awareness of large‑language‑model agents without changing the model weights. By iteratively refining modular skills based on successful and failed trajectories, TRACE raises consistent performance (Pass^3) on the CAR‑bench in‑car assistant tasks from 59.9 % to 94.5 % on GPT‑5.5 and achieves first place on the hidden set with GPT‑5.6‑Sol. The approach demonstrates that a skill‑based, self‑evolution loop can convert a model’s potential into stable, reliable behavior.
By Wenhao Wu, Menghao Zhang, Xin Wang, Zhi Wang, Kun Shao, Jian Luan
arXiv:2607. 28666v1 Announce Type: cross Abstract: Enterprise AI programmes stall at a rate that is widely quoted and poorly explained.
By Prerit Ahuja