The paper introduces Jev, a specialized decision model that selects among given options rather than generating text, and benchmarks it against 19 large language models (LLMs) across 13 multiple-choice tasks in knowledge, reasoning, and multilingual domains. Jev performs competitively with frontier LLMs on knowledge and commonsense benchmarks, achieving the top score on MMLU-Redux and ARC-Challenge, and surpasses most representative and all small LLMs except in mathematical word problems where it lags behind. The study demonstrates that a specialized decision model can match general-purpose LLMs on knowledge-based decisions but struggles with multi-step calculations.
By Xing Li, Qingcheng Chang, Jinzhong Ning, Changfeng Xu, Shenlong Zhang, Yijia Zhang, Ling Luo, Hongfei Lin
InfiFPO is a new preference optimization technique for implicit model fusion that replaces the reference model in Direct Preference Optimization with a fused source model, preserving probability information and avoiding complex vocabulary alignment. By incorporating probability clipping and max‑margin fusion, it aligns a pivot model with human preferences while distilling knowledge from multiple source models. Experiments on 11 benchmarks show that InfiFPO consistently outperforms existing fusion and preference optimization methods, raising Phi‑4’s average score from 79.95 to 83.33 and improving performance in mathematics, coding, and reasoning tasks.
By Yanggan Gu, Yuanyi Wang, Zhaoyi Yan, Yiming Zhang, Qi Zhou, Fei Wu, Hongxia Yang
MedBenchAgent is a multi‑agent framework that automates the entire process of constructing medical vision‑language model benchmarks, from deriving the benchmark specification to generating and auditing evaluation items. It uses a Benchmark Intermediate Representation (BIR) to encode task definitions, evidence mappings, evaluation protocols, and item specifications, separating planning from instantiation. The system achieves a Task‑Space F1 of 90.9% and demonstrates high audit pass rates, outperforming existing task induction methods and showing portability across specialized medical domains.
By Yulin Fu (Beijing University of Posts and Telecommunications), Junren Wang (West China Hospital, Sichuan Provincial Engineering Research Center of Intelligent Diagnosis and Treatment of Breast Diseases), Guangjing Yang (Beijing University of Posts and Telecommunications), Zhangyuan Yu (Beijing University of Posts and Telecommunications), Wanran Sun (Beijing University of Posts and Telecommunications), Jiabao Zhou (Beijing University of Posts and Telecommunications), Jin Yin (West China Hospital, Sichuan Provincial Engineering Research Center of Intelligent Diagnosis and Treatment of Breast Diseases), Qicheng Lao (Beijing University of Posts and Telecommunications)
RIT‑RAG is a retrieval‑augmented generation framework that combines content retrieval with structural navigation of documents. It builds a tree for each document from its table of contents or sitemap, then uses retrieved chunk positions to induce manageable sub‑trees that an LLM agent can navigate, read selectively, and reformulate queries. Experiments on financial, scientific, and customer‑support benchmarks show RIT‑RAG outperforms vanilla, graph‑based, and agentic baselines, achieving up to 11.4‑point accuracy gains on the new EntQABench of 2.84 million technical‑documentation webpages.
By Meghanadh Pulivarthi, Swaraj Kumar Biswal, Kushagra Bhushan, Yatin Nandwani, Sachindra Joshi, Dinesh Raghu
The paper introduces Multi-Agent Self-Supervision (MASS), a recursive self‑improvement technique that alternates between evolutionary workflow optimization and supervised fine‑tuning on self‑generated multi‑agent trajectories. MASS enables a single base model to iteratively propose, execute, and self‑evaluate complex multi‑agent workflows, discovering effective roles and information routing. Experiments with Qwen3.6‑27B show 1.2–1.6× higher performance per output token on four open‑ended benchmarks, and student models trained on MASS traces outperform single‑agent students trained on 1.4× more tokens.
By Hyunin Lee, Jinglue Xu, Jeffrey Seely, Donghyun Lee, Somayeh Sojoudi, Matei Zaharia, Yujin Tang
The paper introduces Stitching Noisy Diffusion Thoughts, a self‑consistency framework that transforms low‑cost diffusion‑sampled reasoning into a reusable pool of step‑level candidates. By sampling many diverse trajectories with a masked diffusion language model, scoring each intermediate step with a process reward model, and stitching the highest‑quality steps across trajectories, the method recomputes only the final answer, separating exploration from evaluation. Experiments on math reasoning benchmarks show that step‑level recombination improves accuracy by up to 23.8% on six tasks while reducing latency by up to 1.8× compared to traditional diffusion models and unified architectures.
By Roy Miles, Aysim Toker, Andreea-Maria Oncescu, Jiankang Deng, Ismail Elezi
The study examines whether large language models truly base legal verdicts on the statutes or precedents they cite. By fixing case facts and swapping the cited authority with an unrelated one, researchers found that while models correctly name the authority 66.7%-100% of the time, the resulting verdict often remains unchanged—only 0.0%-21.7% of the time on CaseHOLD, 30.0%-76.7% on ECHR and SCOTUS, and 43.3%-50.0% on ContractNLI. Even a purpose‑built legal‑reasoning model and larger scales did not close this gap, and adversarial instructions further exposed vulnerabilities in verdict consistency.
By Saisab Sadhu, Shreeyans Arora, Pratinav Seth
The paper presents a memory layer called galahad-kv that stores the key-value state of a large language model’s context blocks (≈16,000 tokens each) on encrypted local NVMe disk, allowing byte‑exact retrieval without recomputation. On a single NVIDIA H100, the authors tested the system with Gemma 4 12B and 31B models over 50 million tokens of real public text, achieving 100 % successful reloads, 2.8×–4.3× faster loading than recomputation, and 8.8×–12.3× lower GPU energy usage while keeping GPU memory flat. The models answered long‑range factual queries with high accuracy (82 % for 12B, 98 % for 31B) and did not hallucinate, demonstrating that stored state can extend effective context without expanding the attention window.
By Sietse Schelpe
ROAR is a Retrieval Opportunity-Aware Refinement framework designed for zero-shot time series forecasting. It enhances predictions by weighting queries according to base-forecast difficulty and the improvement potential of retrieved historical candidates, then learning a gate to control correction strength. The method jointly calibrates the forecasting module and gate, anchoring the final prediction to an initial refined output, and achieves the lowest average MSE across seven benchmarks compared to other methods.
By Jinjin Chi, Lei Feng, Yiming Wang, Lulu Zhang, Yongcheng Jing, Chenyi Li, Ximing Li
MRCert is a masking‑based certified recovery defender that aims to maintain high prediction accuracy while verifying the benignity of labels for adversarially patched inputs in post‑deployment scenarios. It introduces type‑specific necessary properties for both benign and patched samples, enabling a novel label recovery and certification function pair that avoids the accuracy degradation typical of smoothing methods. Experiments show MRCert achieves 35.1% adversarial certified accuracy on ImageNet with 16‑pixel patches, outperforming the state‑of‑the‑art PatchCURE.
By Qilin Zhou, Zhengyuan Wei, Haipeng Wang, Zhuo Wang, Shuo Liu, W. K. Chan
Agent4RE is a self‑refining multi‑agent framework designed for end‑to‑end software requirements engineering (RE). It coordinates specialized agents and includes two iterative improvement loops, and introduces a new real‑world dataset, RE‑E2E, for evaluating RE workflows from elicitation to generation. The framework, along with two enhanced variants that use autonomous self‑refinement or structured human feedback, consistently outperforms baseline prompting methods across eight large language models, achieving higher text‑based metrics and superior human and LLM‑as‑a‑judge ratings.
By Yongjian Tang, Linhan Li, Thomas Runkler
The paper introduces a benchmark for evaluating large language model (LLM) agents in closed‑loop embedded software development. It provides five embedded‑control tasks with realistic build‑and‑runtime feedback, testing agents that must translate requirements into code, self‑verify, and iterate until device behavior meets specifications. Seven GPT‑family and Qwen‑family models were evaluated across four feedback scenarios, revealing that gpt‑5.4 leads in pass rate but does not fully saturate the benchmark, while qwen3.5‑27B is the strongest local model and smaller local models perform poorly.
By Jorge Garc\'ia-Carrasco, Sergio Garc\'ia-Carrasco, Alejandro Mat\'e, Juan Trujillo
The paper introduces SFT-as-context, a training‑free method that lets a parent large language model (LLM) use the response of a supervised fine‑tuned (SFT) model as context to answer queries. This approach preserves the parent model’s general capabilities while acquiring the SFT model’s specialized skills, achieving performance close to the SFT model on fine‑tuned benchmarks and near the parent model on general tasks. Experiments across 19 parent‑SFT pairs and 11 benchmarks show minimal gaps in performance, and the method can solve problems requiring both specialized and general knowledge, even when neither model alone succeeds.
By Kenan Tang, Andong Hua, Chengxuan Qian, Saket Tiwari, Yao Qin
The paper introduces a Spatial Canvas Interface that lets users directly compose poster generation intent in space using semantic, identity, text, and pixel bindings, along with text specifications for elements and overall appearance. It presents Compo, a model adapted from a pretrained image editing system to interpret these canvas inputs, supporting both manual construction and an agentic mode that auto‑generates a canvas from high‑level requests. A scalable training pipeline and benchmark demonstrate that Compo offers stronger compositional control and visual quality than existing general‑purpose and dedicated poster generation models.
By Yitong Wang, Fangyun Wei, Jinjing Zhao, Sirui Zhang, Hongyang Zhang, Dong Chen, Bo Dai, Yan Lu
REACT introduces a rolling-denoising framework that enhances the reactivity of flow-based vision-language-action models while preserving long-horizon context. By maintaining a persistent action buffer and denoising the full horizon at each step, REACT refines executed action blocks across multiple observations, reducing reaction latency and improving trajectory smoothness. Dual decoupling further separates sensing, encoding, denoising, and execution to enable high-frequency updates under realistic compute limits.
By Houlong Xiong, Zhenqi Qiu, Zechen Wang, Suohang Zhang, Yiyu Ren, Wanting Xu, Hongfei Niu, Chengyang He, Ge Sun, Ran Cheng, Qian Zhu
MAST is a motif‑augmented diffusion framework that uses a search tree to jointly elucidate 2D and 3D molecular structures from spectroscopic data. It incorporates explicit motif priors as intermediate evidence during denoising, reducing spectral ambiguity, and casts diffusion sampling as a reward‑guided tree search to prioritize high‑reward trajectories, thereby limiting computational overhead. On the QM9S multi‑spectra benchmark, MAST achieves 94.89% exact recovery, improves 3D fidelity, and maintains high chemical validity and stability.
By Chenghao Jia, Mengdi Liu, Hong Chang, Shiguang Shan, Xilin Chen
VINCIE‑NExT is a unified framework that transfers image editing capabilities to video editing by using in‑context visual demonstrations. It decomposes video editing into a chain of sub‑tasks (Video → Image → Image → Video), routing editing intent through the image domain and enabling joint training from heterogeneous image and video data under a single diffusion objective. The method introduces a novel position encoding to link image demonstrations and video frames, allowing pixel‑faithful appearance propagation, and supports test‑time scaling through progressive diffusion stages.
By Leigang Qu, Feng Cheng, Ziyan Yang, Bangbang Yang, Zhaoyang Huang, Wei Chow, Yicong Li, Wenjie Wang, Tat-Seng Chua, Yan Zeng
MobileWorldBench introduces a benchmark for evaluating vision‑language models as world models in mobile GUI agents, shifting from pixel‑based predictions to natural‑language state transitions. The authors release MobileWorld, a 1.4 million‑sample dataset that enhances VLM world‑modeling performance. They also present a framework that integrates these semantic world models into mobile agent planning, showing improved task success rates.
By Shufan Li, Konstantinos Kallidromitis, Akash Gokul, Yusuke Kato, Kazuki Kozuka, Aditya Grover
MAPF-World introduces an autoregressive action world model for multi‑agent path finding that predicts short‑horizon local observations and neighboring agents’ action intentions, enabling decisions beyond immediate local views. The model incorporates a spatio‑agent positional encoding within Transformer architectures to enhance spatial awareness and agent semantics, fostering coordinated behaviors. Experiments on diverse maps, including a new automated map generator based on real‑world urban layouts, show that MAPF‑World outperforms existing learned solvers, achieving strong performance, robust zero‑shot generalization, and high success rates even at high agent densities.
By Zhanjiang Yang, Yueming Li, Yang Shen, Meng Li, Lijun Sun
The paper introduces OneMillion-Bench (OMB), a benchmark comprising 400 expert-curated tasks across Law, Finance, Industry, Healthcare, and Natural Science. OMB evaluates language agents on complex, real-world scenarios that require authoritative source retrieval, conflict resolution, domain-specific rule application, and constraint decision-making. The benchmark uses a rubric-based scoring system assessing factual accuracy, logical coherence, practical feasibility, and professional compliance to differentiate agent performance on expert-level problems.
By Yang Liu, Jiaqi Li, Jun Bai, Qianyu Yang, Xiaobo Hu, Tao Peng, Zaiyuan Wang, Ran Tian, Jiayun Dong, Chun Zhang, Zixia Jia, Kaiyuan Chen, Yixin Ren, Yang Liu, Yanglihong Xiao, Lingyue Yin, Tiliang Duan, Ge Zhang, Gang Yao, Hao Chen, Yuan Gong, Jianpeng Jiao, Zilong Zheng