QREF is a plug‑and‑play system that improves DeepSearch agents by refining search queries to match the current reasoning intent and filtering evidence to keep only what is relevant. The Query Refiner rewrites queries for complementary coverage across each search batch, while the Evidence Filter removes distracting retrieval results. Experiments show that QREF boosts DeepSearch performance on four benchmarks, achieving significant Pass@1 gains and dramatically reducing ineffective query batches without changing the agents’ architectures or retraining their reasoning policies.
By Yujia Chen, Yuhui Wang, Junzhe Wang, Shihan Dou, Di Liang, Yujiong Shen, Yuhao Zhou, Zhiheng Xi, Ming Zhang, Senjie Jin, Jiayi Chen, Dingwei Zhu, Yajie Yang, Yuan Hua, Chenyao Lu, Tao Gui, Qi Zhang, Xuanjing Huang
DeciCAD introduces a domain‑native harness for CAD agents that defines semantic CAD actions, decision‑oriented observations, environment feedback, and persistent state transitions. Using this harness, the authors develop a learning pipeline that extracts causal decision supervision from interaction trajectories and refines the policy through supervised fine‑tuning and geometry‑grounded online reinforcement learning. Across four CAD benchmarks, DeciCAD outperforms a general‑purpose coding harness, achieving higher F1 and IoU scores while reducing model‑output tokens by 45–88% and improving geometry availability from 66.5% to 73.5%.
By Yufeng Peng, Yixin Zhang, Ziyi Liu, Zhikun Xu, Jiarong Yang, Sijie Wang, Wenjie Feng, Qiyu Kang
HTN‑Orchestrate is a two‑layer framework that separates global task planning from local subtask execution using a hierarchical task network. The Planner builds and updates subtask hierarchies only when prerequisite results are available, while local Orchestrators schedule workers to carry out leaf subtasks. On five benchmarks—including web research, scientific reasoning, and research workflows—the framework outperforms ReAct, Agent Team, and DeepSeek Harness, achieving 89.73% accuracy on BrowseComp and 73.44% on SUPERChem.
By Yanchen Deng, Jianguo Huang, Yao Long Teng, Yixuan Li, Handuo Zhang, Ning Wang, Liangcai Su, Xinyu Wang, Bo An
The paper investigates how environment-agnostic coding agents can be improved without environment-specific engineering. By modifying a Codex-based agent’s instruction to report difficulties and allow automatic continuation, the authors achieve higher mean scores on ARC‑AGI‑3 and AGI Maze, though performance on Terminal‑Bench remains mixed. A scripted controller that handles reset decisions partially recovers the advantage of a specialized agent, demonstrating that effective use of available information and procedures is crucial for these models.
By Sergey Rodionov
A team of four Gemini 4 Argon agents autonomously defeated the Ender Dragon in a sandboxed Minecraft survival world. Using only the open‑source Mineflayer library, the agents coordinated via in‑game chat and shared memory, writing 2,783 programs over 27 hours and 14 minutes while dying 605 times. This marks the first end‑to‑end autonomous completion of Minecraft by AI agents without pre‑scouted seeds or human‑written game‑specific tools.
By Maxence Faldor, Satinder Singh, Roberta Raileanu
The paper reviews how reinforcement learning (RL) can refine diffusion models (DMs) after pretraining, addressing limitations in reasoning, exploration, and alignment. It surveys RL-driven DMs across language, multimodal, robotic, and scientific domains, focusing on trajectory optimization and reward engineering. Experiments highlight the benefits of various RL post‑training strategies and identify key bottlenecks—reward fragility, exploration‑exploitation trade‑offs, and training instability—while proposing a roadmap for future research and a public repository for updates.
By Zhiyuan Ma, Jiabao Wei, Tianyu Wang, Hanqing Wang, Jiadong Pan, Haowen Xu, Liangliang Zhao, Ruixun Liu, Yuhan Wang, Yuwei Zhang, Liang Feng, Junhao Xiao, Kaiyan Zhang, Xuekai Zhu, Daizong Liu, Biqing Qi, Yihao Liu, Ganqu Cui, Ning Ding, Linfeng Zhang, Jianjun Li, Kun He, Xiang Bai, Bowen Zhou
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
TRACER is a training‑free multimodal harness that builds minute‑level action records from continuous video and merges them with clinical data into hourly reports, enabling an LLM agent to query and inspect both modalities. In evaluation, TRACER achieved up to 84.8% accuracy on action detection and 61.0% accuracy on clinician‑informed QA, outperforming the strongest baseline. Across 280 patients and ~60,000 camera‑hours, TRACER detected significantly less purposeful activity in patients with a positive delirium assessment compared to those without (p < 0.0001).
By Zane Durante, Paul Tang, Keshigeyan Chandrasegaran, Amit Kaushal, Shrinidhi Kowshika Lakshmikanth, Feolu Kolawole, Yu Jun Zheng, Yifan Geng, Shobhit Agarwal, Kevin Schulman, Juan Carlos Niebles, Ehsan Adeli, Arnold Milstein, Li Fei-Fei
ARIA is a framework that learns to repair ongoing multi‑agent system executions by turning verified repairs into training data. It builds a large dataset of 3,953 repair pairs from over 24,000 trajectories across multiple frameworks and benchmarks, then trains a compact model to detect when and how to intervene during runtime. The resulting model shows strong repair performance, generalizing to unseen frameworks and host models.
By Bolin Shen, Yushun Dong
MAI-Gym is an open, all-in-one platform that unifies GUI agent evaluation, data collection, and online reinforcement learning training across mobile, computer, and web environments. It supports ten established benchmarks, scales to 20,000 parallel environment instances, and enables automated synthesis of diverse GUI applications and tasks with execution-validated verifiers. Using this platform, the authors trained a 9B-parameter agent, MAI-Gym-9B, achieving state‑of‑the‑art performance on MobileWorld and OSWorld-Verified, and demonstrated that training on a mocked train‑ticketing app improves real‑world task success from 28.8% to 48.1% on a physical phone.
By Panrong Tong, Quyu Kong, Pengxiang Li, Tianyu Xia, Hanzhang Zhou, Lei Wang, Chenglin Cai, Long Li, Long Chen, Jianan Zhang, Zhengwen Zeng, Xu Zhang, Gongjie Zhang, Yue Wang, Steven Hoi
The paper introduces an open midtraining recipe that primes base models for agentic post‑training, filling a gap between pretraining and full fine‑tuning. It trains on 256K context with hundreds of billions of tokens drawn from natural software‑engineering data and a small share of agentic traces. Applying this midtraining to Nemotron 3.5 Lightning and Super models yields measurable gains on SWE‑bench and DeepSWE benchmarks after identical full SFT, and improves robustness across agent harnesses.
By Soumye Singhal, Joseph Jennings, Ka Shun Shum, Alexander Bukharin, Khushi Bhardwaj, Zirui Liu, Venkat Srinivasan, Jonathan Li, Yash Jain, Kezhi Kong, Markus Kliegl, Olivier Delalleau, Sanjeev Satheesh, Mostofa Patwary
The paper proposes a framework for aligning AI agents with deployment requirements through a process called calibration, which avoids the need for source imitation or universal superiority. It introduces three families of requirements—basic capability, technical precision, and user context—alongside a versioned mapping to manage changes. The authors outline four instruments (benchmarks, examples, skills, and role-conditioned cost/utility) that provide tunable reward channels within a fixed acceptance contract, and describe methods for adaptation, qualification, and evaluation, noting that confirmatory validation remains outstanding.
By Yaxiao Liu (PwC China AI Center), Pengbo Liu (PwC China AI Center), Yiwen Liu (PwC China AI Center), Yihua Guan (PwC China AI Center), Jiaxing Song (Tsinghua University), Xingjian Li (PwC China AI Center)
The paper introduces the concept of the effective Speed of Coherent Cognition (S_C^{eff}) as a benchmark for how fast an AI system can adapt to changing reality while maintaining reliable behavior. It defines a challenge where systems must infer reliable actions for increasingly distant future states across a growing observation–reality gap, and tests this using reinforcement‑learning agents trained at a baseline speed of reality (S_R). The study compares Transformer and gated‑MLP baselines with a novel Co^4 architecture inspired by cellular mechanisms of conscious processing, finding that Co^4 achieves a substantially higher S_C^{eff} and thus supports the proposed scaling principle for reliable AI.
By Ahsan Adeel, Talha Bin Riaz
The paper investigates how fine‑tuned language agents can selectively withdraw actions when evidence changes or authority is revoked. Using a synthetic NAQD‑Env environment, the authors compare an original curriculum with an enriched one across multiple system‑prompt framings and evaluate four LoRA checkpoints of Qwen2.5‑3B‑Instruct against instruction‑tuned baselines. Results show that fine‑tuned checkpoints achieve higher decision agreement and withdrawal recall, reduce incorrect withdrawals, and narrow prohibited combinations, though they also omit valid event reports and exhibit mixed preservation of authorized actions.
By Mohamed Abouzahra
FedSkillGit is a federated framework that enables large language model agents to evolve reusable natural‑language skills without sharing raw data. Instead of uploading entire skill documents, clients submit bounded patches and local experience summaries, which are filtered for privacy, deduplicated, and merged into a global skill candidate. The server distributes the candidate for client‑side validation, publishing it only when performance improves, and experiments show it outperforms baselines by up to 22% while approaching centralized results.
By Boxun Lan, Jiamei Li, Weijian Li, Chengyu Yang, Zhuang Qi, Chengcheng Wan, Xiaohong Chen, Ming Hu
The Agent Status Protocol (ASP) is a lightweight specification that enables heterogeneous AI agents to report their operational state to a status server via small JSON events over HTTP POST. The server aggregates these reports and exposes the combined state to observers through Server‑Sent Events (SSE), inferring liveness from missed heartbeats and incorporating observations from supervisors, guardrails, operators, and peers. ASP models state across independent dimensions—lifecycle, health, task, floor, governance—allowing an agent to simultaneously be online, degraded, working on a task, and waiting for the floor, while reusing existing vocabularies from A2A, OFP, and OpenTelemetry GenAI conventions.
By Diego Gosmar, Deborah A. Dahl, Emmett Coin, Csaba Boly\'os, Andreas Zettl
AutoSFT introduces a multi‑agent framework for automated data research in supervised fine‑tuning, where a meta agent separates execution from analysis, gathers diverse proposals, and invokes a synthesis subagent when the data pool is insufficient. The system demonstrates that adding an executor, analysis subagents, and synthesis progressively improves performance on IFEval, IFBench, and MBPP benchmarks, with each component’s impact clearly attributable. These results suggest that effective automation of SFT data research relies on covering a wide range of data actions rather than relying on a single powerful action.
By Bingxiang He, Pengqi Sun, Shihao Peng, Shaoxuan Li, Huiyu Xiao, Hongjian Shi, Zixuan Fu, Ziqing Qiao, Yuxin Zuo, Cheng Qian, Xinqiang Yu, Jiayi Li, Xiusi Chen, Yizhen Song, Xiangyang Wang, Iofai Lei, Yudong Wang, Ning Ding, Zhiyuan Liu, Xiaozhi Wang, Chaojun Xiao
FDE-Gym is a new benchmark that evaluates AI agents on 95 end‑to‑end delivery tasks drawn from real customer scenarios across 18 sectors. Unlike traditional benchmarks that use well‑specified tasks, FDE-Gym places agents in simulated customer environments where they must clarify ambiguous requirements, interact with stakeholders, and uncover unstated constraints to improve a customer’s success metric. In tests, the best model configuration achieved a mean score of 0.56 and solved only 14 of the 95 tasks, highlighting a significant gap between well‑specified problem solving and real‑world AI delivery.
By Dongxu Yang, Hongda Yan, Renxin Wang, Baoli Zhang, Haojia Zuo, Ying Zhao, Youmin Chen, Xiaojian Liao, Congming Gao, Tianyang Jiang, Yixiao Chen, Qing Wang, Zhiguo Yu, Xiaqing Li, Yuhao Zhang, Jiawen Sun, Renzhe Xu, De Mi, Tiancheng Wu, Xianya Fu, Xuanlu Zhao, Bo Feng, Meixuan Ren, Guanming Lin, Ziheng Zhou, Xinhua Wang
The study examines how three context‑engineering techniques—semantic layers, orchestration, and compaction—affect data agents that answer questions over heterogeneous data. Using seven large language models on 1,514 long‑horizon tasks, the authors identify three agent behavior groups (investigators, premature answerers, early stoppers) and show that each technique benefits only the group it aligns with, while sometimes harming others. Combining the techniques yields smaller net gains than the sum of individual effects, leading to guidance on selecting techniques based on agent behavior, task demands, and accuracy‑cost trade‑offs.
By Zhivar Sourati, Marianne Menglin Liu, Zheng Wang, Rongguang Wang, Sai Ashish Somayajula, Tao Sheng, Kyu J. Han, Sujith Ravi, Dan Roth
ReviewPilot is a conversational multi‑agent system designed to automate the PRISMA workflow for systematic literature reviews, covering both paper screening and evidence extraction. Researchers can customize screening criteria and extraction schemas through dialogue, with a project memory preserving configurations and revisions. The system records the rationale for each screening decision and the source of every extracted value, and it has demonstrated superior performance to single‑prompt LLMs and AgentSLR in epidemiology, achieving screening macro F1 scores up to 0.776 and extraction F1 up to 0.788, while also showing close agreement with human annotators in biomedicine and human‑computer interaction case studies.
By Zhaoqian Xue, Shaoze Zhou, Xuexin Li, Junting Mu, Haozheng Liu, Zhongyin Wang, Yichun He, Renkai Ma, Hao Zhou, Sarah Bratt, Siyuan Ma, Chen Chen, Lingyao Li