Daily Papers

Terminal-agent benchmarks have become a primary signal for measuring the coding and system-administration capabilities of large language models. As the market for evaluation environments grows, so does the pressure to ship tasks quickly, often without thorough adversarial review of the verification logic. This paper is a guideline for writing good benchmark tasks, drawn from over a year of contributing to and reviewing tasks for Terminal Bench. Most people write benchmark tasks the way they write prompts. They shouldn't. A prompt is designed to help the agent succeed; a benchmark is designed to find out if it can. We argue that good tasks are adversarial, difficult, and legible, and that a large class of common failure modes -- AI-generated instructions, over-prescriptive specifications, clerical difficulty, oracle solutions that assume hidden knowledge, tests that validate the wrong things, and reward-hackable environments -- are predictable consequences of treating task authoring as prompt authoring. We catalog these failure modes, argue that real difficulty is conceptual rather than environmental, and discuss recent empirical evidence that over 15% of tasks in popular terminal-agent benchmarks are reward-hackable. We hope this serves as a useful reference for benchmark maintainers, task contributors, and researchers using benchmark scores as evidence.

Scalable and reproducible policy evaluation has been a long-standing challenge in robot learning. Evaluations are critical to assess progress and build better policies, but evaluation in the real world, especially at a scale that would provide statistically reliable results, is costly in terms of human time and hard to obtain. Evaluation of increasingly generalist robot policies requires an increasingly diverse repertoire of evaluation environments, making the evaluation bottleneck even more pronounced. To make real-world evaluation of robotic policies more practical, we propose AutoEval, a system to autonomously evaluate generalist robot policies around the clock with minimal human intervention. Users interact with AutoEval by submitting evaluation jobs to the AutoEval queue, much like how software jobs are submitted with a cluster scheduling system, and AutoEval will schedule the policies for evaluation within a framework supplying automatic success detection and automatic scene resets. We show that AutoEval can nearly fully eliminate human involvement in the evaluation process, permitting around the clock evaluations, and the evaluation results correspond closely to ground truth evaluations conducted by hand. To facilitate the evaluation of generalist policies in the robotics community, we provide public access to multiple AutoEval scenes in the popular BridgeData robot setup with WidowX robot arms. In the future, we hope that AutoEval scenes can be set up across institutions to form a diverse and distributed evaluation network.

Robot policy evaluation and deployment remain fragmented by model-specific software dependencies, data representations, and runtime interfaces, so that connecting N policies to M evaluation environments requires O(NM) separate integrations. We present XPolicyLab, a unified standard and open ecosystem that reduces this cost to O(N+M). XPolicyLab specifies common observation, action, and trajectory schemas together with a minimal adapter interface for observation updates, action prediction, batched execution, and episode reset, while a dependency-isolated client/server architecture separates policy inference from environment execution, so that each side retains its native software stack and may run locally or remotely. The ecosystem integrates 42 robot policies and standardizes their installation, debugging, serving, and evaluation workflows. Across these adapters, model-specific code varies by an order of magnitude while the environment-facing loop stays within a few lines of a fixed reference, confirming that the contract confines heterogeneity to the policy side. In a controlled study, conforming to the standard reduces the integration effort of a representative policy from over five hours to two hours, and packaged agent skills reduce it further to thirty minutes. The same adapters serve RoboTwin, RoboDojo simulation, and standardized real-robot evaluation through one interface. XPolicyLab is released as shared infrastructure for reproducible policy comparison and standardized deployment across simulation and physical platforms. Project website: https://xpolicylab.github.io/.

Constructing large-scale datasets for the GitHub issue resolution task is crucial for both training and evaluating the software engineering capabilities of Large Language Models (LLMs). However, the traditional process for creating such benchmarks is notoriously challenging and labor-intensive, particularly in the stages of setting up evaluation environments, grading test outcomes, and validating task instances. In this paper, we propose SWE-Factory, an automated pipeline designed to address these challenges. To tackle these issues, our pipeline integrates three core automated components. First, we introduce SWE-Builder, a multi-agent system that automates evaluation environment construction, which employs four specialized agents that work in a collaborative, iterative loop and leverages an environment memory pool to enhance efficiency. Second, we introduce a standardized, exit-code-based grading method that eliminates the need for manually writing custom parsers. Finally, we automate the fail2pass validation process using these reliable exit code signals. Experiments on 671 issues across four programming languages show that our pipeline can effectively construct valid task instances; for example, with GPT-4.1-mini, our SWE-Builder constructs 269 valid instances at 0.045 per instance, while with Gemini-2.5-flash, it achieves comparable performance at the lowest cost of 0.024 per instance. We also demonstrate that our exit-code-based grading achieves 100% accuracy compared to manual inspection, and our automated fail2pass validation reaches a precision of 0.92 and a recall of 1.00. We hope our automated pipeline will accelerate the collection of large-scale, high-quality GitHub issue resolution datasets for both training and evaluation. Our code and datasets are released at https://github.com/DeepSoftwareAnalytics/swe-factory.

Can Large Language Models (LLMs) simulate human behavior in complex environments? LLMs have recently been shown to exhibit advanced reasoning skills but much of NLP evaluation still relies on static benchmarks. Answering this requires evaluation environments that probe strategic reasoning in competitive, dynamic scenarios that involve long-term planning. We introduce AucArena, a novel simulation environment for evaluating LLMs within auctions, a setting chosen for being highly unpredictable and involving many skills related to resource and risk management, while also being easy to evaluate. We conduct several controlled simulations using state-of-the-art LLMs as bidding agents. We find that through simple prompting, LLMs do indeed demonstrate many of the skills needed for effectively engaging in auctions (e.g., managing budget, adhering to long-term goals and priorities), skills that we find can be sharpened by explicitly encouraging models to be adaptive and observe strategies in past auctions. These results are significant as they show the potential of using LLM agents to model intricate social dynamics, especially in competitive settings. However, we also observe considerable variability in the capabilities of individual LLMs. Notably, even our most advanced models (GPT-4) are occasionally surpassed by heuristic baselines and human agents, highlighting the potential for further improvements in the design of LLM agents and the important role that our simulation environment can play in further testing and refining agent architectures.

In the social world, humans possess the capability to infer and reason about others mental states (such as emotions, beliefs, and intentions), known as the Theory of Mind (ToM). Simultaneously, humans own mental states evolve in response to social situations, a capability we refer to as socialization. Together, these capabilities form the foundation of human social interaction. In the era of artificial intelligence (AI), especially with the development of large language models (LLMs), we raise an intriguing question: How do LLMs perform in terms of ToM and socialization capabilities? And more broadly, can these AI models truly enter and navigate the real social world? Existing research evaluating LLMs ToM and socialization capabilities by positioning LLMs as passive observers from a third person perspective, rather than as active participants. However, compared to the third-person perspective, observing and understanding the world from an egocentric first person perspective is a natural approach for both humans and AI agents. The ToM and socialization capabilities of LLMs from a first person perspective, a crucial attribute for advancing embodied AI agents, remain unexplored. To answer the aforementioned questions and bridge the research gap, we introduce EgoSocialArena, a novel framework designed to evaluate and investigate the ToM and socialization capabilities of LLMs from a first person perspective. It encompasses two evaluation environments: static environment and interactive environment, with seven scenarios: Daily Life, Counterfactual, New World, Blackjack, Number Guessing, and Limit Texas Hold em, totaling 2,195 data entries. With EgoSocialArena, we have conducted a comprehensive evaluation of nine advanced LLMs and observed some key insights regarding the future development of LLMs as well as the capabilities levels of the most advanced LLMs currently available.

Frontier AI safety policies highlight automation of AI research and development (R&D) by AI agents as an important capability to anticipate. However, there exist few evaluations for AI R&D capabilities, and none that are highly realistic and have a direct comparison to human performance. We introduce RE-Bench (Research Engineering Benchmark, v1), which consists of 7 challenging, open-ended ML research engineering environments and data from 71 8-hour attempts by 61 distinct human experts. We confirm that our experts make progress in the environments given 8 hours, with 82% of expert attempts achieving a non-zero score and 24% matching or exceeding our strong reference solutions. We compare humans to several public frontier models through best-of-k with varying time budgets and agent designs, and find that the best AI agents achieve a score 4x higher than human experts when both are given a total time budget of 2 hours per environment. However, humans currently display better returns to increasing time budgets, narrowly exceeding the top AI agent scores given an 8-hour budget, and achieving 2x the score of the top AI agent when both are given 32 total hours (across different attempts). Qualitatively, we find that modern AI agents possess significant expertise in many ML topics -- e.g. an agent wrote a faster custom Triton kernel than any of our human experts' -- and can generate and test solutions over ten times faster than humans, at much lower cost. We open-source the evaluation environments, human expert data, analysis code and agent trajectories to facilitate future research.

Multimodal Large Language Models (MLLMs) exhibit impressive capabilities across a variety of tasks, especially when equipped with carefully designed visual prompts. However, existing studies primarily focus on logical reasoning and visual understanding, while the capability of MLLMs to operate effectively in 3D vision remains an ongoing area of exploration. In this paper, we introduce a novel visual prompting method, called 3DAxisPrompt, to elicit the 3D understanding capabilities of MLLMs in real-world scenes. More specifically, our method leverages the 3D coordinate axis and masks generated from the Segment Anything Model (SAM) to provide explicit geometric priors to MLLMs and then extend their impressive 2D grounding and reasoning ability to real-world 3D scenarios. Besides, we first provide a thorough investigation of the potential visual prompting formats and conclude our findings to reveal the potential and limits of 3D understanding capabilities in GPT-4o, as a representative of MLLMs. Finally, we build evaluation environments with four datasets, i.e., ScanRefer, ScanNet, FMB, and nuScene datasets, covering various 3D tasks. Based on this, we conduct extensive quantitative and qualitative experiments, which demonstrate the effectiveness of the proposed method. Overall, our study reveals that MLLMs, with the help of 3DAxisPrompt, can effectively perceive an object's 3D position in real-world scenarios. Nevertheless, a single prompt engineering approach does not consistently achieve the best outcomes for all 3D tasks. This study highlights the feasibility of leveraging MLLMs for 3D vision grounding/reasoning with prompt engineering techniques.

Trustworthy deployment of LLM-based agents in software systems requires evaluating how they perform on application-specific workflows, with enough granularity to localize where they succeed and fail. Yet existing agent evaluation mechanisms are limited: benchmarks have low construct validity for application-specific workflows and environments, and replica evaluation environments are expensive and prone to drift. We propose Copy-on-Write (CoW) Scoring, a framework that evaluates agent operations directly within application environments using a PostgreSQL-level Copy-on-Write mechanism to isolate agent writes. CoW Scoring produces session- and operation-level scores that highlight where agents' database write operations succeed and fail in a given application environment, enabling inexpensive evaluation and iteration on agent harnesses and tool surfaces. We demonstrate the framework on Plane, an open-source project-management platform, where analysis surfaced specific issues in the tool surface, and corresponding fixes produced measurable improvements on affected models. Python library: https://github.com/trail-ml/agent-cow-python

Scaling robotic foundation models requires diverse training data and reliable evaluation environments. Simulation offers a scalable solution, yet existing generation pipelines remain constrained by predefined assets and skills, a disconnect between scene generation and task generation, and limited support for complex embodiments and physics. We introduce EmbodiedSmith, a framework for scalable embodied data generation through recursive self-improvement (RSI). EmbodiedSmith unifies asset, scene, and task generation in a pipeline that supports autonomous creation and language-driven customization. Its core is an agentic refinement loop: scene generation anticipates downstream task requirements, while task generation guides targeted scene edits, allowing scenes and tasks to iteratively improve one another. This joint refinement improves task generation success, including for long-horizon tasks. The framework further supports mobile manipulators, humanoids, and dexterous hands, as well as interactions involving deformable objects and fluids, broadening the range of behaviors and physical phenomena represented in generated data. Together, these capabilities provide a flexible simulation engine for both robot pretraining and evaluation. Extensive experiments validate the quality, diversity, and generation efficiency of the resulting data, while downstream policy experiments demonstrate that increased data diversity improves generalization.

State-of-the-art visual locomotion controllers are increasingly capable at handling complex visual environments, making evaluating their real-world performance before deployment increasingly difficult. This work intends to narrow this train/evaluation gap by developing a collection of hyper-photo-realistic, closed-loop evaluation environments - The Neverwhere Benchmark Suite - comprised of over sixty 3D Gaussian Splatting reconstructions of urban indoor and outdoor scenes. Our goal is to encourage large-scale and reproducible robot evaluation by making it easier to create and integrate Gaussian splats-based reconstructions into simulated continuous testing setups. We also underscore the potential pitfalls of relying exclusively on 3D Gaussian-generated data for training, by providing policy checkpoints trained over multiple Neverwhere scenes and their performance when evaluated in novel scenes. Our analysis illustrates the necessity of sourcing diverse data to ensure performance. Code and data are available on the project page: https://ziyc.github.io/neverwhere-bench/.

Root cause analysis (RCA) for microservice systems has gained significant attention in recent years. However, there is still no standard benchmark that includes large-scale datasets and supports comprehensive evaluation environments. In this paper, we introduce RCAEval, an open-source benchmark that provides datasets and an evaluation environment for RCA in microservice systems. First, we introduce three comprehensive datasets comprising 735 failure cases collected from three microservice systems, covering various fault types observed in real-world failures. Second, we present a comprehensive evaluation framework that includes fifteen reproducible baselines covering a wide range of RCA approaches, with the ability to evaluate both coarse-grained and fine-grained RCA. We hope that this ready-to-use benchmark will enable researchers and practitioners to conduct extensive analysis and pave the way for robust new solutions for RCA of microservice systems.

Amid the ongoing advancements in Federated Learning (FL), a machine learning paradigm that allows collaborative learning with data privacy protection, personalized FL (pFL)has gained significant prominence as a research direction within the FL domain. Whereas traditional FL (tFL) focuses on jointly learning a global model, pFL aims to balance each client's global and personalized goals in FL settings. To foster the pFL research community, we started and built PFLlib, a comprehensive pFL library with an integrated benchmark platform. In PFLlib, we implemented 37 state-of-the-art FL algorithms (8 tFL algorithms and 29 pFL algorithms) and provided various evaluation environments with three statistically heterogeneous scenarios and 24 datasets. At present, PFLlib has gained more than 1600 stars and 300 forks on GitHub.

Robot imitation data are often multimodal: similar visual-language observations may be followed by different action chunks because human demonstrators act with different short-horizon intents, task phases, or recent context. Existing frame-conditioned VLA policies infer each chunk from the current observation and instruction alone, so under partial observability they may resample different intents across adjacent replanning steps, leading to inter-chunk conflict and unstable execution. We introduce IntentVLA, a history-conditioned VLA framework that encodes recent visual observations into a compact short-horizon intent representation and uses it to condition chunk generation. We further introduce AliasBench, a 12-task ambiguity-aware benchmark on RoboTwin2 with matched training data and evaluation environments that isolate short-horizon observation aliasing. Across AliasBench, SimplerEnv, LIBERO, and RoboCasa, IntentVLA improves rollout stability and outperforms strong VLA baselines

Learning from Observation (LfO) is a fundamental robotic capability that replicates how humans and animals socially learn from each other. Beyond its biological parallels, this modality provides a practical solution for data scaling in sample-inefficient and data-starved domains like robotics. Recent work has demonstrated promising results in learning manipulation skills from human videos, yet progress in this area remains difficult to assess. Existing methods vary widely in assumptions, hardware choices, and environment setups making it difficult to draw meaningful comparisons and identify advances in the field. To address these challenges, we introduce RoboReel: a unified benchmark for evaluating models that learn policies from human videos. RoboReel consists of bundled real-world human demonstration videos, simulated robot trajectories, and evaluation environments on ten manipulation tasks. We develop four test suites to evaluate the models' performance on multiple axes, including the robustness to visual distractors and the ability to complete long-horizon tasks. We further validate RoboReel by recreating a subset of its tasks in the real world, showing that relative performance trends observed in simulation are preserved in the real world. Our benchmark covers learning-from-observation models from different categories, and studies the effectiveness of multiple representation choices in our benchmark evaluation that covers over seven state-of-the-art algorithms (including our VLA based variants) in the field of LfO. Finally, we present an analysis of the different types of algorithms showing that long-horizon tasks and tasks with low tolerances are still challenging for current models. Webpage: https://roboreel.github.io

GUI agent aims to enable automated operations on Mobile/PC devices, which is an important task toward achieving artificial general intelligence. The rapid advancement of VLMs accelerates the development of GUI agents, owing to their powerful capabilities in visual understanding and task planning. However, building a GUI agent remains a challenging task due to the scarcity of operation trajectories, the availability of interactive infrastructure, and the limitation of initial capabilities in foundation models. In this work, we introduce UItron, an open-source foundational model for automatic GUI agents, featuring advanced GUI perception, grounding, and planning capabilities. UItron highlights the necessity of systemic data engineering and interactive infrastructure as foundational components for advancing GUI agent development. It not only systematically studies a series of data engineering strategies to enhance training effects, but also establishes an interactive environment connecting both Mobile and PC devices. In training, UItron adopts supervised finetuning over perception and planning tasks in various GUI scenarios, and then develop a curriculum reinforcement learning framework to enable complex reasoning and exploration for online environments. As a result, UItron achieves superior performance in benchmarks of GUI perception, grounding, and planning. In particular, UItron highlights the interaction proficiency with top-tier Chinese mobile APPs, as we identified a general lack of Chinese capabilities even in state-of-the-art solutions. To this end, we manually collect over one million steps of operation trajectories across the top 100 most popular apps, and build the offline and online agent evaluation environments. Experimental results demonstrate that UItron achieves significant progress in Chinese app scenarios, propelling GUI agents one step closer to real-world application.

LLM agents are shaped not only by their language models, but also by the runtime harness that mediates observation, tool use, action execution, feedback interpretation, and trajectory control. While existing agent adaptation methods mainly update model parameters, many failures in deterministic, rule-governed domains stem from mismatches at the model--environment interface. We propose Life-Harness, a lifecycle-aware runtime harness that improves frozen LLM agents without changing model weights or evaluation environments. Life-Harness evolves from training trajectories by converting recurring interaction failures into reusable interventions across environment contracts, procedural skills, action realization, and trajectory regulation, and remains fixed during held-out evaluation. On seven deterministic environments from τ-bench, τ^2-bench, and AgentBench, Life-Harness improves 116 out of 126 model--environment settings across 18 model backbones, with an average relative improvement of 88.5%. Harnesses evolved only from Qwen3-4B-Instruct trajectories transfer to 17 other models, showing that Life-Harness captures reusable environment-side structure rather than model-specific behavior. These results position runtime interface adaptation as a complementary alternative to model-centric agent training. Code is available at GitHub.

AI has the potential to transform cybersecurity by enabling systems that can autonomously detect, analyze, and remediate software vulnerabilities. However, existing cybersecurity evaluations of AI systems are limited in scale or scope, and fail to capture the end-to-end lifecycle of real-world software vulnerability discovery and remediation. To address this gap, we propose CyberGym-E2E, a large-scale and realistic end-to-end cybersecurity benchmark that comprehensively evaluates AI agents' abilities across the full lifecycle of vulnerability discovery, PoC generation, and patch generation. CyberGym-E2E is comprehensive and scalable, as we build an automated, agent-enhanced pipeline for transforming open-source vulnerability data into realistic evaluation environments. Currently, the benchmark consists of 920 real-world vulnerabilities across 139 different open-source projects.

We introduce Kaggle Game Arena, an open and ever-expanding platform to evaluate large language models (LLMs) through competitive games. Different from static benchmarks, game arena enables models to play head-to-head matchups in structured environments where the gameplay strength naturally increases as models evolve, preventing performance saturation. This technical report details the infrastructure behind Game Arena and describes the three pilot game environments: Chess, Poker, and Werewolf. These environments span perfect information, imperfect information, and multiplayer game settings, enabling a systematic study of models' strategic planning, adaptation, and robustness under uncertainty. For each game, we provide a detailed description of the environment, evaluation metrics, and results from running full competitions across models. Through robust infrastructure and large-scale ground-truth based evaluation, Game Arena ensures reproducibility, transparency and generalizability to new games and variants over time.

Despite recent progress in language models and agents for scientific data-driven discovery, further advancing their capabilities is held back by the absence of verifiable environments representing real-world scientific tasks. To fill this gap, we introduce D3-Gym, the first automatically constructed dataset with verifiable environments for scientific Data-Driven Discovery. D3-Gym comprises (1) 565 tasks sourced from 239 real scientific repositories across four disciplines where (2) each task is equipped with a natural language instruction, an executable environment with pre-installed dependencies, input dataset and artifact previews, a reference code solution, and an automatically synthesized evaluation script. Rigorous evaluation of the quality of the verification signal in D3-Gym confirms that our evaluation scripts achieve 87.5% agreement with human-annotated gold standards and strong alignment in domain-specific evaluation logic, showing their scientific soundness. Further, training on trajectories sampled from D3-Gym yields consistent and substantial gains across Qwen3 models of varying sizes on ScienceAgentBench, boosting Qwen3-32B by 7.8 absolute points and substantially shrinking the gap with strong proprietary models. All D3-Gym artifacts (environments, creation workflow, trajectories, and models) can be found at https://github.com/OSU-NLP-Group/D3-Gym.

Comprehensive, unbiased, and comparable evaluation of modern generalist policies is uniquely challenging: existing approaches for robot benchmarking typically rely on heavy standardization, either by specifying fixed evaluation tasks and environments, or by hosting centralized ''robot challenges'', and do not readily scale to evaluating generalist policies across a broad range of tasks and environments. In this work, we propose RoboArena, a new approach for scalable evaluation of generalist robot policies in the real world. Instead of standardizing evaluations around fixed tasks, environments, or locations, we propose to crowd-source evaluations across a distributed network of evaluators. Importantly, evaluators can freely choose the tasks and environments they evaluate on, enabling easy scaling of diversity, but they are required to perform double-blind evaluations over pairs of policies. Then, by aggregating preference feedback from pairwise comparisons across diverse tasks and environments, we can derive a ranking of policies. We instantiate our approach across a network of evaluators at seven academic institutions using the DROID robot platform. Through more than 600 pairwise real-robot evaluation episodes across seven generalist policies, we demonstrate that our crowd-sourced approach can more accurately rank the performance of existing generalist policies than conventional, centralized evaluation approaches, while being more scalable, resilient, and trustworthy. We open our evaluation network to the community and hope that it can enable more accessible comparisons of generalist robot policies.

Deep reinforcement learning (RL) has shown promising results in robot motion planning with first attempts in human-robot collaboration (HRC). However, a fair comparison of RL approaches in HRC under the constraint of guaranteed safety is yet to be made. We, therefore, present human-robot gym, a benchmark for safe RL in HRC. Our benchmark provides eight challenging, realistic HRC tasks in a modular simulation framework. Most importantly, human-robot gym includes a safety shield that provably guarantees human safety. We are, thereby, the first to provide a benchmark to train RL agents that adhere to the safety specifications of real-world HRC. This bridges a critical gap between theoretic RL research and its real-world deployment. Our evaluation of six environments led to three key results: (a) the diverse nature of the tasks offered by human-robot gym creates a challenging benchmark for state-of-the-art RL methods, (b) incorporating expert knowledge in the RL training in the form of an action-based reward can outperform the expert, and (c) our agents negligibly overfit to training data.

We introduce AInsteinBench, a large-scale benchmark for evaluating whether large language model (LLM) agents can operate as scientific computing development agents within real research software ecosystems. Unlike existing scientific reasoning benchmarks which focus on conceptual knowledge, or software engineering benchmarks that emphasize generic feature implementation and issue resolving, AInsteinBench evaluates models in end-to-end scientific development settings grounded in production-grade scientific repositories. The benchmark consists of tasks derived from maintainer-authored pull requests across six widely used scientific codebases, spanning quantum chemistry, quantum computing, molecular dynamics, numerical relativity, fluid dynamics, and cheminformatics. All benchmark tasks are carefully curated through multi-stage filtering and expert review to ensure scientific challenge, adequate test coverage, and well-calibrated difficulty. By leveraging evaluation in executable environments, scientifically meaningful failure modes, and test-driven verification, AInsteinBench measures a model's ability to move beyond surface-level code generation toward the core competencies required for computational scientific research.

Multi-Agent Path Finding (MAPF) remains a fundamental challenge in robotics, where classical centralized approaches exhibit exponential growth in joint-state complexity as the number of agents increases. This paper investigates Quadratic Unconstrained Binary Optimization (QUBO) as a structurally scalable alternative for simultaneous multi-robot path planning. This approach is a robotics-oriented QUBO formulation incorporating BFS-based logical pre-processing (achieving over 95% variable reduction), adaptive penalty design for collision and constraint enforcement, and a time-windowed decomposition strategy that enables execution within current hardware limitations. An experimental evaluation in grid environments with up to four robots demonstrated near-optimal solutions in dense scenarios and favorable scaling behavior compared to sequential classical planning. These results establish a practical and reproducible baseline for future quantum and quantum-inspired multi-robot coordinations.

Modern forest monitoring workflows increasingly benefit from the growing availability of high-resolution satellite imagery and advances in deep learning. Two persistent challenges in this context are accurate pixel-level change detection and meaningful semantic change captioning for complex forest dynamics. While large language models (LLMs) are being adapted for interactive data exploration, their integration with vision-language models (VLMs) for remote sensing image change interpretation (RSICI) remains underexplored. To address this gap, we introduce an LLM-driven agent for integrated forest change analysis that supports natural language querying across multiple RSICI tasks. The proposed system builds upon a multi-level change interpretation (MCI) vision-language backbone with LLM-based orchestration. To facilitate adaptation and evaluation in forest environments, we further introduce the Forest-Change dataset, which comprises bi-temporal satellite imagery, pixel-level change masks, and multi-granularity semantic change captions generated using a combination of human annotation and rule-based methods. Experimental results show that the proposed system achieves mIoU and BLEU-4 scores of 67.10% and 40.17% on the Forest-Change dataset, and 88.13% and 34.41% on LEVIR-MCI-Trees, a tree-focused subset of LEVIR-MCI benchmark for joint change detection and captioning. These results highlight the potential of interactive, LLM-driven RSICI systems to improve accessibility, interpretability, and efficiency of forest change analysis. All data and code are publicly available at https://github.com/JamesBrockUoB/ForestChat.

The increasing availability of high-resolution satellite imagery, together with advances in deep learning, creates new opportunities for enhancing forest monitoring workflows. Two central challenges in this domain are pixel-level change detection and semantic change interpretation, particularly for complex forest dynamics. While large language models (LLMs) are increasingly adopted for data exploration, their integration with vision-language models (VLMs) for remote sensing image change interpretation (RSICI) remains underexplored, especially beyond urban environments. We introduce Forest-Chat, an LLM-driven agent designed for integrated forest change analysis. The proposed framework enables natural language querying and supports multiple RSICI tasks, including change detection, change captioning, object counting, deforestation percentage estimation, and change reasoning. Forest-Chat builds upon a multi-level change interpretation (MCI) vision-language backbone with LLM-based orchestration, and incorporates zero-shot change detection via a foundation change detection model together with an interactive point-prompt interface to support fine-grained user guidance. To facilitate adaptation and evaluation in forest environments, we introduce the Forest-Change dataset, comprising bi-temporal satellite imagery, pixel-level change masks, and multi-granularity semantic change captions generated through a combination of human annotation and rule-based methods. Experimental results demonstrate that Forest-Chat achieves strong performance on Forest-Change and on LEVIR-MCI-Trees, a tree-focused subset of LEVIR-MCI, for joint change detection and captioning, highlighting the potential of interactive, LLM-driven RSICI systems to improve accessibility, interpretability, and analytical efficiency in forest change analysis.

Recent agent benchmarks increasingly ground evaluation in executable environments, from code repair to web navigation, app APIs, and function calling. Yet completing consequential work beyond code requires more than producing a plausible response or valid tool call: agents must gather missing information over multiple turns, follow domain policies, coordinate dependent tools, and realize the correct persistent state transition without collateral effects. In this paper, we introduce Thinkingbox, a sandbox for tool-agent-user interaction that provides isolated MCP-compatible tool sessions, complete execution traces, and outcome evaluation over terminal backend state. Built on this sandbox, Thinkingbox-bench contains 507 policy-conditioned workflows across numerous scenarios, including retail, hospitality, auto insurance, neobank internal IT, and consulting IT/HR support. Each attempt is evaluated by task-specific executable checks that accept valid trajectories while rejecting wrong, missing, or extra effects; designated tasks additionally check required properties of the final response. Across proprietary and open-weight models, the strongest achieves 65.36% pass@1, but only 25.25% pass^20. Moreover, many failed trials show clean termination and valid state-changing actions, showing that response or tool-call-level signals are not clear proxies for end-to-end task completion. Thinkingbox-bench reveals a large gap between occasionally finding a successful trajectory and reliably completing stateful business tasks. We release both Thinkingbox and Thinkingbox-Bench: https://github.com/microsoft/thinkingbox

As agentic systems increasingly rely on reinforcement learning from verifiable rewards, standardized ``gym'' infrastructure has become essential for rapid iteration, reproducibility, and fair comparison. Vision agents lack such infrastructure, limiting systematic study of what drives their learning and where current models fall short. We introduce Gym-V, a unified platform of 179 procedurally generated visual environments across 10 domains with controllable difficulty, enabling controlled experiments that were previously infeasible across fragmented toolkits. Using it, we find that observation scaffolding is more decisive for training success than the choice of RL algorithm, with captions and game rules determining whether learning succeeds at all. Cross-domain transfer experiments further show that training on diverse task categories generalizes broadly while narrow training can cause negative transfer, with multi-turn interaction amplifying all of these effects. Gym-V is released as a convenient foundation for training environments and evaluation toolkits, aiming to accelerate future research on agentic VLMs.

Graphical User Interface (GUI) agents, driven by Multi-modal Large Language Models (MLLMs), have emerged as a promising paradigm for enabling intelligent interaction with digital systems. This paper provides a structured survey of recent advances in GUI agents, focusing on architectures enhanced by Reinforcement Learning (RL). We first formalize GUI agent tasks as Markov Decision Processes and discuss typical execution environments and evaluation metrics. We then review the modular architecture of (M)LLM-based GUI agents, covering Perception, Planning, and Acting modules, and trace their evolution through representative works. Furthermore, we categorize GUI agent training methodologies into Prompt-based, Supervised Fine-Tuning (SFT)-based, and RL-based approaches, highlighting the progression from simple prompt engineering to dynamic policy learning via RL. Our summary illustrates how recent innovations in multimodal perception, decision reasoning, and adaptive action generation have significantly improved the generalization and robustness of GUI agents in complex real-world environments. We conclude by identifying key challenges and future directions for building more capable and reliable GUI agents.

The Boreas Road Trip (Boreas-RT) dataset extends the multi-season Boreas dataset to new and diverse locations that pose challenges for modern autonomous driving algorithms. Boreas-RT comprises 60 sequences collected over 9 real-world routes, totalling 643 km of driving. Each route is traversed multiple times, enabling evaluation in identical environments under varying traffic and, in some cases, weather conditions. The data collection platform includes a 5MP FLIR Blackfly S camera, a 360 degree Navtech RAS6 Doppler-enabled spinning radar, a 128-channel 360 degree Velodyne Alpha Prime lidar, an Aeva Aeries II FMCW Doppler-enabled lidar, a Silicon Sensing DMU41 inertial measurement unit, and a Dynapar wheel encoder. Centimetre-level ground truth is provided via post-processed Applanix POS LV GNSS-INS data. The dataset includes precise extrinsic and intrinsic calibrations, a publicly available development kit, and a live leaderboard for odometry and metric localization. Benchmark results show that many state-of-the-art odometry and localization algorithms overfit to simple driving environments and degrade significantly on the more challenging Boreas-RT routes. Boreas-RT provides a unified dataset for evaluating multi-modal algorithms across diverse road conditions. The dataset, leaderboard, and development kit are available at www.boreas.utias.utoronto.ca.

Digital Adoption Platforms (DAPs) are embedded overlays widely used on web systems to guide users through operations inside a page, helping them get started with unfamiliar interfaces quickly. Completing a real task, however, rarely means clicking a few buttons on a single page: it takes a sequence of actions that unfolds across changing page states. Prior studies have also treated automated web agent actions and guide text generation as two separate problems, and most of them feed models textual page representations such as the DOM or accessibility trees rather than the rendered screens that humans actually operate on. In this work we introduce MAG, the first benchmark that unifies task execution and guide writing into a single Multimodal Action and Guide task, with two grounding schemes over screenshots: Set-of-Mark element selection and raw pixel coordinates. We further build a complete harness for this compound task, covering annotation with LLM assistance and human verification, training, evaluation in live environments, and joint metrics for actions and guides. With this harness we evaluate frontier API models and open multimodal models, and report detailed analyses. Finally, we design a GRPO training method augmented with expert trajectories, which nearly doubles the success rate of a supervised 9B agent (from 6.9% to 13.2%) and improves guide quality at the same time. Even the strongest model completes fewer than 40% of the tasks, leaving ample room for future research.

We present INTELLECT-3, a 106B-parameter Mixture-of-Experts model (12B active) trained with large-scale reinforcement learning on our end-to-end RL infrastructure stack. INTELLECT-3 achieves state of the art performance for its size across math, code, science and reasoning benchmarks, outperforming many larger frontier models. We open-source the model together with the full infrastructure stack used to create it, including RL frameworks, complete recipe, and a wide collection of environments, built with the verifiers library, for training and evaluation from our Environments Hub community platform. Built for this effort, we introduce prime-rl, an open framework for large-scale asynchronous reinforcement learning, which scales seamlessly from a single node to thousands of GPUs, and is tailored for agentic RL with first-class support for multi-turn interactions and tool use. Using this stack, we run both SFT and RL training on top of the GLM-4.5-Air-Base model, scaling RL training up to 512 H200s with high training efficiency.

In the realm of robot action recognition, identifying distinct but spatially proximate arm movements using vision systems in noisy environments poses a significant challenge. This paper studies robot arm action recognition in noisy environments using machine learning techniques. Specifically, a vision system is used to track the robot's movements followed by a deep learning model to extract the arm's key points. Through a comparative analysis of machine learning methods, the effectiveness and robustness of this model are assessed in noisy environments. A case study was conducted using the Tic-Tac-Toe game in a 3-by-3 grid environment, where the focus is to accurately identify the actions of the arms in selecting specific locations within this constrained environment. Experimental results show that our approach can achieve precise key point detection and action classification despite the addition of noise and uncertainties to the dataset.

Large language model (LLM) agents integrated with external tools are vulnerable to indirect prompt injections embedded in environmental states. However, existing studies largely rely on manually implemented or reused environments, stochastic LLM-based tool simulation, and predefined injection locations, limiting scalable security research across broader domains. To bridge this gap, we propose **ToolHazard**, a scalable adversarial environment synthesis framework that reduces human engineering and supports expansion with additional seed domains and compute. Through an Environment Simulator, an Attacker Agent, and a User Simulator, ToolHazard synthesizes executable stateful environments, discovers viable injection points and generates environment-specific payloads, and constructs state-grounded long-horizon tasks. Based on ToolHazard, we build **ToolHazard-Bench** for stress-testing agents under complex workflows and diverse environmental attacks. Experiments reveal substantial agent vulnerabilities and show that injection timing and placement affect attack effectiveness. Moreover, ToolHazard-generated alignment data improves security on both ToolHazard-Bench and AgentDojo while preserving benign task utility.

We investigate whether GPT-6-Astra, a general-purpose foundation model, can navigate unfamiliar environments using its own perception, reasoning, and decision-making capabilities. Our evaluation focuses on zero-shot vision-and-language navigation in continuous environments (VLN-CE) through a minimal interface in the Codex harness, aiming to unleash GPT-6-Astra's full potential for navigation. Using monocular RGB, GPT-6-Astra decides when to observe, how to move, and when to stop, without navigation-specific fine-tuning, a trained waypoint predictor, or a pre-built scene map. Our evaluation yields four key findings and implications. First, GPT-6-Astra achieves strong zero-shot navigation performance using only monocular RGB observations. On R2R-CE-100, GPT-6-Astra (ultra reasoning) achieves a success rate of 81.3%, exceeding the strongest reported zero-shot and even train-based success rates by 15.3 and 9.2 percentage points, respectively. Second, GPT-6-Astra exhibits promising capabilities in interpreting multi-stage instructions, understanding the environment, and adjusting routes. Third, execution and goal-verification failures persist even with ultra reasoning. Fourth, these results motivate combining general-purpose model capabilities with navigation-specific expertise. Based on these findings, future VLN research should investigate which aspects of instruction interpretation, spatial understanding, and navigation decision-making general-purpose models can handle directly, and where navigation-specific learning can extend their capabilities. This includes exploring how spatial representations, navigation experience, and learned control skills can improve progress tracking, error recovery, and goal verification while preserving the flexibility to adjust routes.

As Large Language Models (LLMs) evolve into autonomous agents, existing safety evaluations face a fundamental trade-off: manual benchmarks are costly, while LLM-based simulators are scalable but suffer from logic hallucination. We present AutoControl Arena, an automated framework for frontier AI risk evaluation built on the principle of logic-narrative decoupling. By grounding deterministic state in executable code while delegating generative dynamics to LLMs, we mitigate hallucination while maintaining flexibility. This principle, instantiated through a three-agent framework, achieves over 98% end-to-end success and 60% human preference over existing simulators. To elicit latent risks, we vary environmental Stress and Temptation across X-Bench (70 scenarios, 7 risk categories). Evaluating 9 frontier models reveals: (1) Alignment Illusion: risk rates surge from 21.7% to 54.5% under pressure, with capable models showing disproportionately larger increases; (2) Scenario-Specific Safety Scaling: advanced reasoning improves robustness for direct harms but worsens it for gaming scenarios; and (3) Divergent Misalignment Patterns: weaker models cause non-malicious harm while stronger models develop strategic concealment.

Coding agents increasingly run inside organizations whose security controls (scoped credentials, restricted egress, read-only filesystems, non-root execution) constrain them like any other software. Existing benchmarks, however, evaluate agents almost exclusively in permissive sandboxes, so it is unknown how performance changes when policy is enforced. In this work, we evaluate 12 coding agents on Terminal-Bench 2.1 across nested security policy levels derived from common real-world enterprise restrictions. Hardening is never free but far from uniform: under the strictest policy, success losses reach 18.3 points and cost inflation 167.3\%, and the two axes disagree; the model that best preserves success is also the one that loses the most efficiency, so model choice is policy-dependent. Beyond aggregate scores, we characterize how agents behave when policy blocks their actions and decompose the failures hardening induces: runs grind into timeouts or wrong solutions rather than stopping early, in a mix that differs by model. To ground comparisons, we verify task solvability under the strictest policy, separating model failures from tasks the policy forecloses. We release Boundary-Bench, an open-source hardening plugin enabling policy-constrained evaluation of coding agents on Terminal-Bench and compatible benchmarks.

Despite recent progress in using Large Language Models (LLMs) for automatically generating 3D scenes, generated scenes often lack realistic spatial layouts and object attributes found in real-world environments. As this problem stems from insufficiently detailed, coarse-grained instructions, advancing 3D scene synthesis guided by more detailed, fine-grained instructions that reflect real-world environments becomes crucial. Without such realistic scenes, training embodied agents in unrealistic environments can lead them to learn priors that diverge significantly from real-world physics and semantics, degrading their performance when deployed. Thus, verifying the alignment between the fine-grained instruction and the generated scene is essential for effective learning. However, current evaluation methods, such as CLIPScore and vision-language models (VLMs), often fail to reliably assess such alignment. This shortcoming arises primarily from their shallow understanding of 3D scenes, which often leads to improperly grounded scene components. To address this, we introduce LEGO-Eval, an evaluation framework equipped with diverse tools designed to explicitly ground scene components, enabling more accurate alignment assessments. We also present LEGO-Bench, a benchmark of detailed instructions that specify complex layouts and attributes of real-world environments. Experiments demonstrate that LEGO-Eval outperforms VLM-as-a-judge by 0.41 F1 score in assessing scene-instruction alignment. Benchmarking with LEGO-Bench reveals significant limitations in current generation methods. Across all evaluated approaches, success rates reached at most 10% in generating scenes that fully align with fine-grained instructions.

RTK-SLAM systems integrate simultaneous localization and mapping (SLAM) with real-time kinematic (RTK) GNSS positioning, promising both relative consistency and globally referenced coordinates for efficient georeferenced surveying. A critical and underappreciated issue is that the standard evaluation metric, Absolute Trajectory Error (ATE), first fits an optimal rigid-body transformation between the estimated trajectory and reference before computing errors. This so-called SE(3) alignment absorbs global drift and systematic errors, making trajectories appear more accurate than they are in practice, and is unsuitable for evaluating the global accuracy of RTK-SLAM. We present a geodetically referenced dataset and evaluation methodology that expose this gap. A key design principle is that the RTK receiver is used solely as a system input, while ground truth is established independently via a geodetic total station. This separation is absent from all existing datasets, where GNSS typically serves as (part of) the ground truth. The dataset is collected with a handheld RTK-SLAM device, comprising two scenes. We evaluate LiDAR-inertial, visual-inertial, and LiDAR-visual-inertial RTK-SLAM systems alongside standalone RTK, reporting direct global accuracy and SE(3)-aligned relative accuracy to make the gap explicit. Results show that SE(3) alignment can underestimate absolute positioning error by up to 76\%. RTK-SLAM achieves centimeter-level absolute accuracy in open-sky conditions and maintains decimeter-level global accuracy indoors, where standalone RTK degrades to tens of meters. The dataset, calibration files, and evaluation scripts are publicly available at https://rtk-slam-dataset.github.io/.

This study systematically performed an extensive real-world evaluation of the performances of all configurations of YOLOv8, YOLOv9, YOLOv10, YOLO11( or YOLOv11), and YOLOv12 object detection algorithms in terms of precision, recall, mean Average Precision at 50\% Intersection over Union (mAP@50), and computational speeds including pre-processing, inference, and post-processing times immature green apple (or fruitlet) detection in commercial orchards. Additionally, this research performed and validated in-field counting of the fruitlets using an iPhone and machine vision sensors. Among the configurations, YOLOv12l recorded the highest recall rate at 0.90, compared to all other configurations of YOLO models. Likewise, YOLOv10x achieved the highest precision score of 0.908, while YOLOv9 Gelan-c attained a precision of 0.903. Analysis of mAP@0.50 revealed that YOLOv9 Gelan-base and YOLOv9 Gelan-e reached peak scores of 0.935, with YOLO11s and YOLOv12l following closely at 0.933 and 0.931, respectively. For counting validation using images captured with an iPhone 14 Pro, the YOLO11n configuration demonstrated outstanding accuracy, recording RMSE values of 4.51 for Honeycrisp, 4.59 for Cosmic Crisp, 4.83 for Scilate, and 4.96 for Scifresh; corresponding MAE values were 4.07, 3.98, 7.73, and 3.85. Similar performance trends were observed with RGB-D sensor data. Moreover, sensor-specific training on Intel Realsense data significantly enhanced model performance. YOLOv11n achieved highest inference speed of 2.4 ms, outperforming YOLOv8n (4.1 ms), YOLOv9 Gelan-s (11.5 ms), YOLOv10n (5.5 ms), and YOLOv12n (4.6 ms), underscoring its suitability for real-time object detection applications. (YOLOv12 architecture, YOLOv11 Architecture, YOLOv12 object detection, YOLOv11 object detecion, YOLOv12 segmentation)

Online reinforcement learning (RL) with verifiable rewards has driven rapid progress in mathematics, coding, and tool use, but scaling this success to GUI agents remains challenging due to the high cost of interactive environments and the scarcity of realistic training tasks. We introduce MAI-Gym, an open all-in-one platform that unifies GUI agent evaluation, data collection, and online RL training across mobile, computer, and web environments. Through a shared rollout pipeline, MAI-Gym ensures consistent agent behavior across evaluation and training, supports ten established benchmarks, and scales to 20,000 parallel environment instances. Beyond scalable infrastructure, MAI-Gym enables automated synthesis of diverse GUI applications and challenging tasks, including resettable mocked apps and coherent cross-app worlds with execution-validated task verifiers. Using MAI-Gym, we train MAI-Gym-9B, a unified 9B agent for mobile and computer use, achieving state-of-the-art performance among open GUI agents under 10B parameters: 78.4% on MobileWorld and 76.2% on OSWorld-Verified. Importantly, RL in a mocked train-ticketing app transfers to its real-world counterpart, improving task success on a phys- ical phone from 28.8% to 48.1%. We release the framework, synthesis skills, mocked apps, 2,000 verified tasks, and the model checkpoint to support repro- ducible research and further advances in scalable GUI agent training.

Large language models are shifting from passive information providers to active agents intended for complex workflows. However, their deployment as reliable AI workers in enterprise is stalled by benchmarks that fail to capture the intricacies of professional environments, specifically, the need for long-horizon planning amidst persistent state changes and strict access protocols. In this work, we introduce EnterpriseOps-Gym, a benchmark designed to evaluate agentic planning in realistic enterprise settings. Specifically, EnterpriseOps-Gym features a containerized sandbox with 164 database tables and 512 functional tools to mimic real-world search friction. Within this environment, agents are evaluated on 1,150 expert-curated tasks across eight mission-critical verticals (including Customer Service, HR, and IT). Our evaluation of 14 frontier models reveals critical limitations in state-of-the-art models: the top-performing Claude Opus 4.5 achieves only 37.4% success. Further analysis shows that providing oracle human plans improves performance by 14-35 percentage points, pinpointing strategic reasoning as the primary bottleneck. Additionally, agents frequently fail to refuse infeasible tasks (best model achieves 53.9%), leading to unintended and potentially harmful side effects. Our findings underscore that current agents are not yet ready for autonomous enterprise deployment. More broadly, EnterpriseOps-Gym provides a concrete testbed to advance the robustness of agentic planning in professional workflows.

Data agents use code to process heterogeneous files and derive analytical insights. Scaling their training faces costly environment curation, privacy and intellectual-property restrictions, and difficulty verifying open-ended findings without known generative truth. Our core insight is to specify the world behind the files before generating the files themselves. We introduce Data World, a multi-agent engine that synthesizes environments, tasks, and rubrics from explicit mechanisms. Independent audits check target recoverability from observed files; rubrics support evaluation and RL rewards. Human preference-based curation for question realism selects 100 tasks across 50 held-out worlds and 16 domains for DataWorld-Bench. The engine also yields 2,135 training tasks from 1,037 disjoint worlds. Evaluation of ten frontier models shows that the benchmark remains challenging, particularly for insight discovery, with a consistent computation-insight gap. RL-zero training on the synthesized tasks produces DataWorld-35B-A3B from Qwen3.6-35B-A3B without trajectory distillation. Under matched evaluation, it improves over its base by 7.06 points on DataWorld-Bench, 14.28 on DAComp, 20.90 on DABstep-Hard, and 12.13 on JobBench, a broader occupational benchmark.

Large language models (LLMs) are advancing at an unprecedented pace globally, with regions increasingly adopting these models for applications in their primary language. Evaluation of these models in diverse linguistic environments, especially in low-resource languages, has become a major challenge for academia and industry. Existing evaluation frameworks are disproportionately focused on English and a handful of high-resource languages, thereby overlooking the realistic performance of LLMs in multilingual and lower-resource scenarios. To address this gap, we introduce GlotEval, a lightweight framework designed for massively multilingual evaluation. Supporting seven key tasks (machine translation, text classification, summarization, open-ended generation, reading comprehension, sequence labeling, and intrinsic evaluation), spanning over dozens to hundreds of languages, GlotEval highlights consistent multilingual benchmarking, language-specific prompt templates, and non-English-centric machine translation. This enables a precise diagnosis of model strengths and weaknesses in diverse linguistic contexts. A multilingual translation case study demonstrates GlotEval's applicability for multilingual and language-specific evaluations.

Deploying robots in human-shared environments requires a deep understanding of how nearby agents and objects interact. Employing causal inference to model cause-and-effect relationships facilitates the prediction of human behaviours and enables the anticipation of robot interventions. However, a significant challenge arises due to the absence of implementation of existing causal discovery methods within the ROS ecosystem, the standard de-facto framework in robotics, hindering effective utilisation on real robots. To bridge this gap, in our previous work we proposed ROS-Causal, a ROS-based framework designed for onboard data collection and causal discovery in human-robot spatial interactions. In this work, we present an experimental evaluation of ROS-Causal both in simulation and on a new dataset of human-robot spatial interactions in a lab scenario, to assess its performance and effectiveness. Our analysis demonstrates the efficacy of this approach, showcasing how causal models can be extracted directly onboard by robots during data collection. The online causal models generated from the simulation are consistent with those from lab experiments. These findings can help researchers to enhance the performance of robotic systems in shared environments, firstly by studying the causal relations between variables in simulation without real people, and then facilitating the actual robot deployment in real human environments. ROS-Causal: https://lcastri.github.io/roscausal

AI agents deployed as persistent assistants must maintain correct beliefs as their information environment evolves. In practice, evidence is scattered across heterogeneous sources that often contradict one another, new information can invalidate earlier conclusions, and user preferences surface through corrections rather than explicit instructions. Existing benchmarks largely assume static, single-authority settings and do not evaluate whether agents can keep up with this complexity. We introduce ClawArena, a benchmark for evaluating AI agents in evolving information environments. Each scenario maintains a complete hidden ground truth while exposing the agent only to noisy, partial, and sometimes contradictory traces across multi-channel sessions, workspace files, and staged updates. Evaluation is organized around three coupled challenges: multi-source conflict reasoning, dynamic belief revision, and implicit personalization, whose interactions yield a 14-category question taxonomy. Two question formats, multi-choice (set-selection) and shell-based executable checks, test both reasoning and workspace grounding. The current release contains 64 scenarios across 8 professional domains, totaling 1{,}879 evaluation rounds and 365 dynamic updates. Experiments on five agent frameworks and five language models show that both model capability (15.4% range) and framework design (9.2%) substantially affect performance, that self-evolving skill frameworks can partially close model-capability gaps, and that belief revision difficulty is determined by update design strategy rather than the mere presence of updates. Code is available at https://github.com/aiming-lab/ClawArena.

Aligning autonomous agents with human intent remains a central challenge in modern AI. A key manifestation of this challenge is reward hacking, whereby agents appear successful under the evaluation signal while violating the intended objective. Reward hacking has been observed across a wide range of settings, yet methods for reliably measuring it at scale remain lacking. In this work, we introduce a new evaluation paradigm for measuring reward hacking. Whereas prior studies have primarily analyzed it post hoc by inspecting agent trajectories, we instead embed detectable reward hacking opportunities directly into environments. This makes their exploitation verifiable by design, enabling deterministic and automated measurement of whether and how agents exploit such vulnerabilities. We instantiate this approach in TextArena and release Hack-Verifiable TextArena, a testbed in which reward hacking can be measured reliably. Using this benchmark, we analyze reward hacking behavior across language models in diverse environments and settings. We open source the code at https://github.com/MajoRoth/hack-verifiable-environments/.

Video generation models aspire to simulate dynamic environments, and several benchmarks now evaluate memory consistency across frames. However, most assess consistency only while the target remains in view, and the few that force objects out of view evaluate static scenes where nothing changes during occlusion. To bridge this gap, we introduce MemoBench, a diagnostic benchmark built around the disappear-and-reappear paradigm in dynamically changing environments: a target object undergoes a physical process, disappears from view, and must be correctly recovered in its updated state upon reappearance. We curate 360 ground-truth clips spanning synthetic and real-world scenes, and design an evaluation suite combining automated metrics with VQA-based assessment across four diagnostic pillars. Evaluation of eight state-of-the-art models reveals key insights and open challenges regarding memory consistency under the disappear-and-reappear paradigm.

We introduce Forecast-Dojo, a replayable environment for benchmarking and training LLM forecasting agents. It combines resolved prediction-market questions with dated news, allowing agents to research an event and revisit their predictions at successive historical dates. The same tasks and tools support repeated evaluation, collection of training interactions, and feedback from recorded outcomes without waiting for new events to resolve. Forecast-Dojo contains 1,568 Polymarket events, split by time into training and evaluation periods, and 18.8M dated news articles. In an evaluation of 12 models, research tools lower Brier score for all 12. Forecasts also improve as events unfold, with the largest gains at steps where more newly dated evidence is recorded. Every model still trails historical market forecasts in both Brier score and accuracy. A belief notebook carried between dates lowers research cost but does not consistently improve forecast quality. Beyond evaluation, Forecast-Dojo provides interaction trajectories and outcome feedback for agent learning, with supervised fine-tuning as a proof of concept.

Recent advancements in Multimodal Large Language Models (MLLMs) have led to significant improvements across various multimodal benchmarks. However, as evaluations shift from static datasets to open-world, dynamic environments, current game-based benchmarks remain inadequate because they lack visual-centric tasks and fail to assess the diverse reasoning skills required for real-world decision-making. To address this, we introduce Visual-centric Multiple Abilities Game Evaluation (V-MAGE), a game-based evaluation framework designed to assess visual reasoning capabilities of MLLMs. V-MAGE features five diverse games with 30+ handcrafted levels, testing models on core visual skills such as positioning, trajectory tracking, timing, and visual memory, alongside higher-level reasoning like long-term planning and deliberation. We use V-MAGE to evaluate leading MLLMs, revealing significant challenges in their visual perception and reasoning. In all game environments, the top-performing MLLMs, as determined by Elo rating comparisons, exhibit a substantial performance gap compared to humans. Our findings highlight critical limitations, including various types of perceptual errors made by the models, and suggest potential avenues for improvement from an agent-centric perspective, such as refining agent strategies and addressing perceptual inaccuracies. Code is available at https://github.com/CSU-JPG/V-MAGE.

Training generalist policies for robotic manipulation has shown great promise, as they enable language-conditioned, multi-task behaviors across diverse scenarios. However, evaluating these policies remains difficult because real-world testing is expensive, time-consuming, and labor-intensive. It also requires frequent environment resets and carries safety risks when deploying unproven policies on physical robots. Manually creating and populating simulation environments with assets for robotic manipulation has not addressed these issues, primarily due to the significant engineering effort required and the substantial sim-to-real gap, both in terms of physics and rendering. In this paper, we explore the use of action-conditional video generation models as a scalable way to learn world models for policy evaluation. We demonstrate how to incorporate action conditioning into existing pre-trained video generation models. This allows leveraging internet-scale in-the-wild online videos during the pre-training stage and alleviates the need for a large dataset of paired video-action data, which is expensive to collect for robotic manipulation. Our paper examines the effect of dataset diversity, pre-trained weights, and common failure cases for the proposed evaluation pipeline. Our experiments demonstrate that across various metrics, including policy ranking and the correlation between actual policy values and predicted policy values, these models offer a promising approach for evaluating policies without requiring real-world interactions.

We introduce Gaia2, a benchmark for evaluating large language model agents in realistic, asynchronous environments. Unlike prior static or synchronous evaluations, Gaia2 introduces scenarios where environments evolve independently of agent actions, requiring agents to operate under temporal constraints, adapt to noisy and dynamic events, resolve ambiguity, and collaborate with other agents. Each scenario is paired with a write-action verifier, enabling fine-grained, action-level evaluation and making Gaia2 directly usable for reinforcement learning from verifiable rewards. Our evaluation of state-of-the-art proprietary and open-source models shows that no model dominates across capabilities: GPT-5 (high) reaches the strongest overall score of 42% pass@1 but fails on time-sensitive tasks, Claude-4 Sonnet trades accuracy and speed for cost, Kimi-K2 leads among open-source models with 21% pass@1. These results highlight fundamental trade-offs between reasoning, efficiency, robustness, and expose challenges in closing the "sim2real" gap. Gaia2 is built on a consumer environment with the open-source Agents Research Environments platform and designed to be easy to extend. By releasing Gaia2 alongside the foundational ARE framework, we aim to provide the community with a flexible infrastructure for developing, benchmarking, and training the next generation of practical agent systems.

Vision-Language Navigation in Continuous Environments (VLNCE), where an agent follows instructions and moves freely to reach a destination, is a key research problem in embodied AI. However, most existing approaches are sensitive to viewpoint changes, i.e. variations in camera height and viewing angle. Here we introduce a more general scenario, V^2-VLNCE (VLNCE with Varied Viewpoints) and propose a view-invariant post-training framework, called VIL (View Invariant Learning), that makes existing navigation policies more robust to changes in camera viewpoint. VIL employs a contrastive learning framework to learn sparse and view-invariant features. We also introduce a teacher-student framework for the Waypoint Predictor Module, a standard part of VLNCE baselines, where a view-dependent teacher model distills knowledge into a view-invariant student model. We employ an end-to-end training paradigm to jointly optimize these components. Empirical results show that our method outperforms state-of-the-art approaches on V^2-VLNCE by 8-15\% measured on Success Rate for two standard benchmark datasets R2R-CE and RxR-CE. Evaluation of VIL in standard VLNCE settings shows that despite being trained for varied viewpoints, VIL often still improves performance. On the harder RxR-CE dataset, our method also achieved state-of-the-art performance across all metrics. This suggests that adding VIL does not diminish the standard viewpoint performance and can serve as a plug-and-play post-training method. We further evaluate VIL for simulated camera placements derived from real robot configurations (e.g. Stretch RE-1, LoCoBot), showing consistent improvements of performance. Finally, we present a proof-of-concept real-robot evaluation in two physical environments using a panoramic RGB sensor combined with LiDAR. The code is available at https://github.com/realjoshqsun/V2-VLNCE.

We study decision making in environments where the reward is only partially observed, but can be modeled as a function of an action and an observed context. This setting, known as contextual bandits, encompasses a wide variety of applications including health-care policy and Internet advertising. A central task is evaluation of a new policy given historic data consisting of contexts, actions and received rewards. The key challenge is that the past data typically does not faithfully represent proportions of actions taken by a new policy. Previous approaches rely either on models of rewards or models of the past policy. The former are plagued by a large bias whereas the latter have a large variance. In this work, we leverage the strength and overcome the weaknesses of the two approaches by applying the doubly robust technique to the problems of policy evaluation and optimization. We prove that this approach yields accurate value estimates when we have either a good (but not necessarily consistent) model of rewards or a good (but not necessarily consistent) model of past policy. Extensive empirical comparison demonstrates that the doubly robust approach uniformly improves over existing techniques, achieving both lower variance in value estimation and better policies. As such, we expect the doubly robust approach to become common practice.

This paper introduces an open-source benchmark for evaluating Vision-Language Models (VLMs) on Optical Character Recognition (OCR) tasks in dynamic video environments. We present a curated dataset containing 1,477 manually annotated frames spanning diverse domains, including code editors, news broadcasts, YouTube videos, and advertisements. Three state of the art VLMs - Claude-3, Gemini-1.5, and GPT-4o are benchmarked against traditional OCR systems such as EasyOCR and RapidOCR. Evaluation metrics include Word Error Rate (WER), Character Error Rate (CER), and Accuracy. Our results highlight the strengths and limitations of VLMs in video-based OCR tasks, demonstrating their potential to outperform conventional OCR models in many scenarios. However, challenges such as hallucinations, content security policies, and sensitivity to occluded or stylized text remain. The dataset and benchmarking framework are publicly available to foster further research.

The rapid proliferation of LLM-based autonomous agents in real operating system environments introduces a new category of safety risk beyond content safety: behavior jailbreak, where an adversary induces an agent to execute dangerous OS-level operations with irreversible consequences. Existing benchmarks either evaluate safety at the semantic layer alone, missing physical-layer harms, or fail to isolate test cases, letting earlier runs contaminate later ones. We present LITMUS (LLM-agents In-OS Testing for Measuring Unsafe Subversion), a benchmark addressing both gaps via a semantic-physical dual verification mechanism and OS-level state rollback. LITMUS comprises 819 high-risk test cases organized into one harmful seed subset and six attack-extended subsets covering three adversarial paradigms (jailbreak speaking, skill injection, and entity wrapping), plus a fully automated multi-agent evaluation framework judging behavior at both conversational and OS-level physical layers. Evaluation across frontier agents reveals three findings: (1) current agents lack effective safety awareness, with strong models (e.g., Claude Sonnet 4.6) still executing 40.64% of high-risk operations; (2) agents exhibit pervasive Execution Hallucination (EH), verbally refusing a request while the dangerous operation has already completed at the system level, invisible to every prior semantic-only framework; and (3) skill injection and entity wrapping attacks achieve high success rates, exposing pronounced agent vulnerabilities. LITMUS provides the first standardized platform for reproducible, physically grounded behavioral safety evaluation of LLM agents in real OS environments.

Skillful mobile operation in three-dimensional environments is a primary topic of study in Artificial Intelligence. The past two years have seen a surge of creative work on navigation. This creative output has produced a plethora of sometimes incompatible task definitions and evaluation protocols. To coordinate ongoing and future research in this area, we have convened a working group to study empirical methodology in navigation research. The present document summarizes the consensus recommendations of this working group. We discuss different problem statements and the role of generalization, present evaluation measures, and provide standard scenarios that can be used for benchmarking.

Apollo did not reach the Moon merely because its engineers could solve difficult equations. It succeeded by turning a distant ambition into a mission architecture of explicit objectives, simulation, verification, and repeated correction. AI now faces a similar transition: frontier models can solve difficult tasks once the problem, tools, and success criteria are specified, yet consequential real-world challenges rarely arrive in an executable or verifiable form. We introduce Apodex Discovery, a framework for building and evaluating discoverative AI through the heavy-duty solver, a system comprising a foundation model, harness, tools, and control policies that pursues extended, stateful, verifiable investigations. It has three core components. First, a problem-scouting process surveyed 561 industries across 16 sectors, assembled 423 high-value real-world problems, and selected 20 for the initial release. Second, a common environment-task-episode abstraction provides data, tools, constraints, feedback, trajectory recording, and verification of intermediate artifacts and final submissions. Third, HDS6 evaluates Tools, Repair, Alternatives, Coherence, Evidence, and Scope independently of final-task success. In AAV capsid design, Apodex surpassed the published state of the art by 7% across viability, tropism, structure prediction, and generative design. In drug repurposing and reformulation, a task-specific biomedical environment improved the mean normalized prediction score of GPT-5.5 and GPT-5.6-sol by 2.5 and 7.6 points over the same closed-book backbone. Controlled ablations show that the fixed TRACES episode interface enables attribution of performance differences to specific solver components. Apodex Discovery moves AI evaluation beyond predefined benchmarks toward verifiable investigations aimed at genuine discovery.

Game engines provide real-time simulation, rendering, physics, interaction, networking, and asset pipelines, making them valuable not only for games but also for 3D applications in healthcare, robotics, architecture, manufacturing, and related domains. Because game development is where these systems are most mature and publicly available, it offers a practical testbed for evaluating coding agents that must modify C++ code within stateful, interactive, real-time systems. We present GameEngineBench, a benchmark for evaluating coding agents on scoped C++ implementation tasks inside Unreal Engine 5 projects, built from nine real-world game repositories. The evaluation set consists of 110 tasks spanning gameplay mechanics, multiplayer behavior, AI and world orchestration, animation and movement, UI and session code, loading behavior, online-service integration, persistence, data serialization, XR behavior, and rendering-oriented plugins. These tasks require models to make native C++ changes that compile and satisfy behavioral tests within executable Unreal Engine projects. Across twelve evaluated configurations, the strongest model reaches 55.5\% pass@1, while 31 tasks remain unsolved by every configuration. Our results demonstrate that frontier coding agents continue to struggle with deeply integrated C++ development for real-time interactive software, highlighting game-engine benchmarks as a valuable complement to existing software engineering evaluations.

Computer-use agents (CUAs) have advanced along two separate lines: graphical interaction and software development through code and the command line. Real digital work requires both, interleaved rather than stacked end to end. We study hybrid CUAs that autonomously decide when to explore an interface, implement software, and run and visually verify their artifacts. We introduce RecreationWorld, a five-platform framework built around recreation: given a running reference, an agent must discover its behavior and build a faithful implementation with no prescribed workflow. RecreationWorld provides reproducible environments on Ubuntu, macOS, Windows, Android, and Web, plus a unified harness with native GUI control and coding tools. The running reference serves as an oracle for hidden behavioral tests, providing execution-grounded rewards. We scale trajectory generation with high-quality open-source applications. Models trained on these trajectories improve across five out-of-distribution coding and hybrid computer-use benchmarks and more frequently verify their rendered outputs, providing evidence of transfer beyond recreation. For held-out evaluation, we introduce RecreationBench, comprising 250 diverse tasks across domains and platforms. Reference-grounded programmatic and visual assertions cover action-conditioned outcomes at multiple interaction depths; each is validated on the reference and by human reviewers before the suite is frozen for automatic scoring. GPT-6 Astra leads at 58.1% overall, but passes all programmatic tests on just 2.8% of tasks. Agents reproduce static interface structure more reliably than interactions and computed outputs, while generated applications remain smaller and more monolithic than their references. We release the benchmark, environments, and test suites.

Autonomous agents have made rapid progress in general-purpose computer use, but reliable automation of professional industrial engineering remains out of reach, as engineering workflows demand reasoning over geometric and physical constraints and dependencies preserved across software and design stages. We present EngiWorld, the first benchmark structured around the complete design loop: 1,301 expert-curated tasks spanning 6 engineering domains (CAD, CAE, CAM, BIM, EDA, and 3D visualization) and 26 professional software platforms, with both GUI and CLI interfaces and 6 task types ranging from software-selection to open-ended tasks. We further introduce an artifact-centric evaluation methodology built on a unified domain-verifier suite, which programmatically checks the geometric validity, physical feasibility, and rule compliance of final and intermediate artifacts, and scores quantitative design tasks continuously by specification attainment rather than binary success. Evaluation of seven frontier models reveals a substantial capability gap: the strongest model achieves an EngiScore of only 44.3, and just 3.6% of multi-software attempts succeed. EngiWorld provides the first rigorous foundation for measuring progress toward agents that operate professional engineering software end to end.

3D simulated environments play a critical role in Embodied AI, but their creation requires expertise and extensive manual effort, restricting their diversity and scope. To mitigate this limitation, we present Holodeck, a system that generates 3D environments to match a user-supplied prompt fully automatedly. Holodeck can generate diverse scenes, e.g., arcades, spas, and museums, adjust the designs for styles, and can capture the semantics of complex queries such as "apartment for a researcher with a cat" and "office of a professor who is a fan of Star Wars". Holodeck leverages a large language model (GPT-4) for common sense knowledge about what the scene might look like and uses a large collection of 3D assets from Objaverse to populate the scene with diverse objects. To address the challenge of positioning objects correctly, we prompt GPT-4 to generate spatial relational constraints between objects and then optimize the layout to satisfy those constraints. Our large-scale human evaluation shows that annotators prefer Holodeck over manually designed procedural baselines in residential scenes and that Holodeck can produce high-quality outputs for diverse scene types. We also demonstrate an exciting application of Holodeck in Embodied AI, training agents to navigate in novel scenes like music rooms and daycares without human-constructed data, which is a significant step forward in developing general-purpose embodied agents.

The integration of Large Language Models (LLMs) into Development Environments (IDEs) has become a focal point in modern software development. LLMs such as OpenAI GPT-3.5/4 and Code Llama offer the potential to significantly augment developer productivity by serving as intelligent, chat-driven programming assistants. However, utilizing LLMs out of the box is unlikely to be optimal for any given scenario. Rather, each system requires the LLM to be honed to its set of heuristics to ensure the best performance. In this paper, we introduce the Copilot evaluation harness: a set of data and tools for evaluating LLM-guided IDE interactions, covering various programming scenarios and languages. We propose our metrics as a more robust and information-dense evaluation than previous state of the art evaluation systems. We design and compute both static and execution based success metrics for scenarios encompassing a wide range of developer tasks, including code generation from natural language (generate), documentation generation from code (doc), test case generation (test), bug-fixing (fix), and workspace understanding and query resolution (workspace). These success metrics are designed to evaluate the performance of LLMs within a given IDE and its respective parameter space. Our learnings from evaluating three common LLMs using these metrics can inform the development and validation of future scenarios in LLM guided IDEs.

Autonomous agents that accomplish complex computer tasks with minimal human interventions have the potential to transform human-computer interaction, significantly enhancing accessibility and productivity. However, existing benchmarks either lack an interactive environment or are limited to environments specific to certain applications or domains, failing to reflect the diverse and complex nature of real-world computer use, thereby limiting the scope of tasks and agent scalability. To address this issue, we introduce OSWorld, the first-of-its-kind scalable, real computer environment for multimodal agents, supporting task setup, execution-based evaluation, and interactive learning across various operating systems such as Ubuntu, Windows, and macOS. OSWorld can serve as a unified, integrated computer environment for assessing open-ended computer tasks that involve arbitrary applications. Building upon OSWorld, we create a benchmark of 369 computer tasks involving real web and desktop apps in open domains, OS file I/O, and workflows spanning multiple applications. Each task example is derived from real-world computer use cases and includes a detailed initial state setup configuration and a custom execution-based evaluation script for reliable, reproducible evaluation. Extensive evaluation of state-of-the-art LLM/VLM-based agents on OSWorld reveals significant deficiencies in their ability to serve as computer assistants. While humans can accomplish over 72.36% of the tasks, the best model achieves only 12.24% success, primarily struggling with GUI grounding and operational knowledge. Comprehensive analysis using OSWorld provides valuable insights for developing multimodal generalist agents that were not possible with previous benchmarks. Our code, environment, baseline models, and data are publicly available at https://os-world.github.io.

Crop detection is integral for precision agriculture applications such as automated yield estimation or fruit picking. However, crop detection, e.g., apple detection in orchard environments remains challenging due to a lack of large-scale datasets and the small relative size of the crops in the image. In this work, we address these challenges by reformulating the apple detection task in a semi-supervised manner. To this end, we provide the large, high-resolution dataset MAD comprising 105 labeled images with 14,667 annotated apple instances and 4,440 unlabeled images. Utilizing this dataset, we also propose a novel Semi-Supervised Small Apple Detection system S^3AD based on contextual attention and selective tiling to improve the challenging detection of small apples, while limiting the computational overhead. We conduct an extensive evaluation on MAD and the MSU dataset, showing that S^3AD substantially outperforms strong fully-supervised baselines, including several small object detection systems, by up to 14.9%. Additionally, we exploit the detailed annotations of our dataset w.r.t. apple properties to analyze the influence of relative size or level of occlusion on the results of various systems, quantifying current challenges.

For web agents to be practically useful, they must adapt to the continuously evolving web environment characterized by frequent updates to user interfaces and content. However, most existing benchmarks only capture the static aspects of the web. To bridge this gap, we introduce WebCanvas, an innovative online evaluation framework for web agents that effectively addresses the dynamic nature of web interactions. WebCanvas contains three main components to facilitate realistic assessments: (1) A novel evaluation metric which reliably capture critical intermediate actions or states necessary for task completions while disregarding noise caused by insignificant events or changed web-elements. (2) A benchmark dataset called Mind2Web-Live, a refined version of original Mind2Web static dataset containing 542 tasks with 2439 intermediate evaluation states; (3) Lightweight and generalizable annotation tools and testing pipelines that enables the community to collect and maintain the high-quality, up-to-date dataset. Building on WebCanvas, we open-source an agent framework with extensible modules for reasoning, providing a foundation for the community to conduct online inference and evaluations. Our best-performing agent achieves a task success rate of 23.1% and a task completion rate of 48.8% on the Mind2Web-Live test set. Additionally, we analyze the performance discrepancies across various websites, domains, and experimental environments. We encourage the community to contribute further insights on online agent evaluation, thereby advancing this field of research.

Today's video-conferencing tools support a rich range of professional and social activities, but their generic, grid-based environments cannot be easily adapted to meet the varying needs of distributed collaborators. To enable end-user customization, we developed BlendScape, a system for meeting participants to compose video-conferencing environments tailored to their collaboration context by leveraging AI image generation techniques. BlendScape supports flexible representations of task spaces by blending users' physical or virtual backgrounds into unified environments and implements multimodal interaction techniques to steer the generation. Through an evaluation with 15 end-users, we investigated their customization preferences for work and social scenarios. Participants could rapidly express their design intentions with BlendScape and envisioned using the system to structure collaboration in future meetings, but experienced challenges with preventing distracting elements. We implement scenarios to demonstrate BlendScape's expressiveness in supporting distributed collaboration techniques from prior work and propose composition techniques to improve the quality of environments.

Meta-reinforcement learning algorithms can enable robots to acquire new skills much more quickly, by leveraging prior experience to learn how to learn. However, much of the current research on meta-reinforcement learning focuses on task distributions that are very narrow. For example, a commonly used meta-reinforcement learning benchmark uses different running velocities for a simulated robot as different tasks. When policies are meta-trained on such narrow task distributions, they cannot possibly generalize to more quickly acquire entirely new tasks. Therefore, if the aim of these methods is to enable faster acquisition of entirely new behaviors, we must evaluate them on task distributions that are sufficiently broad to enable generalization to new behaviors. In this paper, we propose an open-source simulated benchmark for meta-reinforcement learning and multi-task learning consisting of 50 distinct robotic manipulation tasks. Our aim is to make it possible to develop algorithms that generalize to accelerate the acquisition of entirely new, held-out tasks. We evaluate 7 state-of-the-art meta-reinforcement learning and multi-task learning algorithms on these tasks. Surprisingly, while each task and its variations (e.g., with different object positions) can be learned with reasonable success, these algorithms struggle to learn with multiple tasks at the same time, even with as few as ten distinct training tasks. Our analysis and open-source environments pave the way for future research in multi-task learning and meta-learning that can enable meaningful generalization, thereby unlocking the full potential of these methods.

We introduce Reasoning Gym (RG), a library of reasoning environments for reinforcement learning with verifiable rewards. It provides over 100 data generators and verifiers spanning multiple domains including algebra, arithmetic, computation, cognition, geometry, graph theory, logic, and various common games. Its key innovation is the ability to generate virtually infinite training data with adjustable complexity, unlike most previous reasoning datasets, which are typically fixed. This procedural generation approach allows for continuous evaluation across varying difficulty levels. Our experimental results demonstrate the efficacy of RG in both evaluating and reinforcement learning of reasoning models.

The widespread deployment of Large Language Model (LLM) agents across real-world applications has unlocked tremendous potential, while raising some safety concerns. Among these concerns, the self-replication risk of LLM agents driven by objective misalignment (just like Agent Smith in the movie The Matrix) has drawn growing attention. Previous studies mainly examine whether LLM agents can self-replicate when directly instructed, potentially overlooking the risk of spontaneous replication driven by real-world settings (e.g., ensuring survival against termination threats). In this paper, we present a comprehensive evaluation framework for quantifying self-replication risks. Our framework establishes authentic production environments and realistic tasks (e.g., dynamic load balancing) to enable scenario-driven assessment of agent behaviors. Designing tasks that might induce misalignment between users' and agents' objectives makes it possible to decouple replication success from risk and capture self-replication risks arising from these misalignment settings. We further introduce Overuse Rate (OR) and Aggregate Overuse Count (AOC) metrics, which precisely capture the frequency and severity of uncontrolled replication. In our evaluation of 21 state-of-the-art open-source and proprietary models, we observe that over 50\% of LLM agents display a pronounced tendency toward uncontrolled self-replication, reaching an overall Risk Score (Phi_R) above a safety threshold of 0.5 when subjected to operational pressures. Our results underscore the urgent need for scenario-driven risk assessment and robust safeguards in the practical deployment of LLM agents.

Can generative agents be trusted in multimodal environments? Despite advances in large language and vision-language models that enable agents to act autonomously and pursue goals in rich settings, their ability to reason about safety, coherence, and trust across modalities remains limited. We introduce a reproducible simulation framework for evaluating agents along three dimensions: (1) safety improvement over time, including iterative plan revisions in text-visual scenarios; (2) detection of unsafe activities across multiple categories of social situations; and (3) social dynamics, measured as interaction counts and acceptance ratios of social exchanges. Agents are equipped with layered memory, dynamic planning, multimodal perception, and are instrumented with SocialMetrics, a suite of behavioral and structural metrics that quantifies plan revisions, unsafe-to-safe conversions, and information diffusion across networks. Experiments show that while agents can detect direct multimodal contradictions, they often fail to align local revisions with global safety, reaching only a 55 percent success rate in correcting unsafe plans. Across eight simulation runs with three models - Claude, GPT-4o mini, and Qwen-VL - five agents achieved average unsafe-to-safe conversion rates of 75, 55, and 58 percent, respectively. Overall performance ranged from 20 percent in multi-risk scenarios with GPT-4o mini to 98 percent in localized contexts such as fire/heat with Claude. Notably, 45 percent of unsafe actions were accepted when paired with misleading visuals, showing a strong tendency to overtrust images. These findings expose critical limitations in current architectures and provide a reproducible platform for studying multimodal safety, coherence, and social dynamics.

We introduce the Few-Shot Object Learning (FewSOL) dataset for object recognition with a few images per object. We captured 336 real-world objects with 9 RGB-D images per object from different views. Object segmentation masks, object poses and object attributes are provided. In addition, synthetic images generated using 330 3D object models are used to augment the dataset. We investigated (i) few-shot object classification and (ii) joint object segmentation and few-shot classification with the state-of-the-art methods for few-shot learning and meta-learning using our dataset. The evaluation results show that there is still a large margin to be improved for few-shot object classification in robotic environments. Our dataset can be used to study a set of few-shot object recognition problems such as classification, detection and segmentation, shape reconstruction, pose estimation, keypoint correspondences and attribute recognition. The dataset and code are available at https://irvlutd.github.io/FewSOL.

Graphical user interface (GUI) agents can substantially improve productivity by automating frequently executed long-latency tasks on mobile devices. However, existing evaluation benchmarks are still constrained to limited applications, simple tasks, and coarse-grained metrics. To address this, we introduce AndroidLens, a challenging evaluation framework for mobile GUI agents, comprising 571 long-latency tasks in both Chinese and English environments, each requiring an average of more than 26 steps to complete. The framework features: (1) tasks derived from real-world user scenarios across 38 domains, covering complex types such as multi-constraint, multi-goal, and domain-specific tasks; (2) static evaluation that preserves real-world anomalies and allows multiple valid paths to reduce bias; and (3) dynamic evaluation that employs a milestone-based scheme for fine-grained progress measurement via Average Task Progress (ATP). Our evaluation indicates that even the best models reach only a 12.7% task success rate and 50.47% ATP. We also underscore key challenges in real-world environments, including environmental anomalies, adaptive exploration, and long-term memory retention.

Environments are the bottleneck for self-improving agents. Current terminal benchmarks were built for evaluation, not training; reinforcement learning requires a scalable pipeline, not just a dataset. We introduce Endless Terminals, a fully autonomous pipeline that procedurally generates terminal-use tasks without human annotation. The pipeline has four stages: generating diverse task descriptions, building and validating containerized environments, producing completion tests, and filtering for solvability. From this pipeline we obtain 3255 tasks spanning file operations, log management, data processing, scripting, and database operations. We train agents using vanilla PPO with binary episode level rewards and a minimal interaction loop: no retrieval, multi-agent coordination, or specialized tools. Despite this simplicity, models trained on Endless Terminals show substantial gains: on our held-out dev set, Llama-3.2-3B improves from 4.0% to 18.2%, Qwen2.5-7B from 10.7% to 53.3%, and Qwen3-8B-openthinker-sft from 42.6% to 59.0%. These improvements transfer to human-curated benchmarks: models trained on Endless Terminals show substantial gains on held out human curated benchmarks: on TerminalBench 2.0, Llama-3.2-3B improves from 0.0% to 2.2%, Qwen2.5-7B from 2.2% to 3.4%, and Qwen3-8B-openthinker-sft from 1.1% to 6.7%, in each case outperforming alternative approaches including models with more complex agentic scaffolds. These results demonstrate that simple RL succeeds when environments scale.

World Generation Models are emerging as a cornerstone of next-generation multimodal intelligence systems. Unlike traditional 2D visual generation, World Models aim to construct realistic, dynamic, and physically consistent 3D/4D worlds from images, videos, or text. These models not only need to produce high-fidelity visual content but also maintain coherence across space, time, physics, and instruction control, enabling applications in virtual reality, autonomous driving, embodied intelligence, and content creation. However, prior benchmarks emphasize different evaluation dimensions and lack a unified assessment of world-realism capability. To systematically evaluate World Models, we introduce the 4DWorldBench, which measures models across four key dimensions: Perceptual Quality, Condition-4D Alignment, Physical Realism, and 4D Consistency. The benchmark covers tasks such as Image-to-3D/4D, Video-to-4D, Text-to-3D/4D. Beyond these, we innovatively introduce adaptive conditioning across multiple modalities, which not only integrates but also extends traditional evaluation paradigms. To accommodate different modality-conditioned inputs, we map all modality conditions into a unified textual space during evaluation, and further integrate LLM-as-judge, MLLM-as-judge, and traditional network-based methods. This unified and adaptive design enables more comprehensive and consistent evaluation of alignment, physical realism, and cross-modal coherence. Preliminary human studies further demonstrate that our adaptive tool selection achieves closer agreement with subjective human judgments. We hope this benchmark will serve as a foundation for objective comparisons and improvements, accelerating the transition from "visual generation" to "world generation." Our project can be found at https://yeppp27.github.io/4DWorldBench.github.io/.

Recent advances in autonomous LLM agents demonstrate their ability to improve performance through iterative interaction with the environment. We define this paradigm as Test-Time Improvement (TTI). However, the mechanisms under how and why TTI succeed or fail remain poorly understood, and existing evaluation metrics fail to capture their task optimization efficiency, behavior adaptation after erroneous actions, and the specific utility of working memory for task completion. To address these gaps, we propose Test-time Improvement Diagnostic Evaluation (TIDE), an agent-agnostic and environment-agnostic framework that decomposes TTI into three comprehensive and interconnected dimensions. The framework measures (1) the overall temporal dynamics of task completion and (2) identifies whether performance is primarily constrained by recursive looping behaviors or (3) by burdensome accumulated memory. Through extensive experiments across diverse agents and environments, TIDE highlights that improving agent performance requires more than scaling internal reasoning, calling for explicitly optimizing the interaction dynamics between the agent and the environment.

Recent advancements in image generation models have enabled the prediction of future Graphical User Interface (GUI) states based on user instructions. However, existing benchmarks primarily focus on general domain visual fidelity, leaving the evaluation of state transitions and temporal coherence in GUI-specific contexts underexplored. To address this gap, we introduce GEBench, a comprehensive benchmark for evaluating dynamic interaction and temporal coherence in GUI generation. GEBench comprises 700 carefully curated samples spanning five task categories, covering both single-step interactions and multi-step trajectories across real-world and fictional scenarios, as well as grounding point localization. To support systematic evaluation, we propose GE-Score, a novel five-dimensional metric that assesses Goal Achievement, Interaction Logic, Content Consistency, UI Plausibility, and Visual Quality. Extensive evaluations on current models indicate that while they perform well on single-step transitions, they struggle significantly with maintaining temporal coherence and spatial grounding over longer interaction sequences. Our findings identify icon interpretation, text rendering, and localization precision as critical bottlenecks. This work provides a foundation for systematic assessment and suggests promising directions for future research toward building high-fidelity generative GUI environments. The code is available at: https://github.com/stepfun-ai/GEBench.

The promise of general-purpose agents - systems that perform tasks in unfamiliar environments without domain-specific engineering - remains largely unrealized. Existing agents are predominantly specialized, and while emerging implementations like OpenAI SDK Agent and Claude Code hint at broader capabilities, no systematic evaluation of their general performance has been pursued. Current agentic benchmarks assume domain-specific integration, encoding task information in ways that preclude fair evaluation of general agents. This paper frames general-agent evaluation as a first-class research objective. We propose conceptual principles for such evaluation, a Unified Protocol enabling agent-benchmark integration, and Exgentic - a practical framework for general agent evaluation. We benchmark five prominent agent implementations across six environments as the first Open General Agent Leaderboard. Our experiments show that general agents generalize across diverse environments, achieving performance comparable to domain-specific agents without any environment-specific tuning. We release our evaluation protocol, framework, and leaderboard to establish a foundation for systematic research on general-purpose agents.

The advancement of large language model (LLM) based agents has shifted AI evaluation from single-turn response assessment to multi-step task completion in interactive environments. We present an empirical study evaluating frontier AI models on 150 workplace tasks within a realistic e-commerce RL environment from Surge. Our analysis reveals an empirically-derived hierarchy of agentic capabilities that models must master for real-world deployment: (1) tool use, (2) planning and goal formation, (3) adaptability, (4) groundedness, and (5) common-sense reasoning. Even the best-performing models fail approximately 40\% of the tasks, with failures clustering predictably along this hierarchy. Weaker models struggle with fundamental tool use and planning, whereas stronger models primarily fail on tasks requiring contextual inference beyond explicit instructions. We introduce a task-centric design methodology for RL environments that emphasizes diversity and domain expert contributions, provide detailed failure analysis, and discuss implications for agent development. Our findings suggest that while current frontier models can demonstrate coherent multi-step behavior, substantial capability gaps remain before achieving human-level task completion in realistic workplace settings.

Evaluating robotics policies across thousands of environments and thousands of tasks is infeasible with existing approaches. This motivates the need for a new methodology for scalable robotics policy evaluation. In this paper, we propose dWorldEval, which uses a discrete diffusion world model as a scalable evaluation proxy for robotics policies. Specifically, dWorldEval maps all modalities - including vision, language, and robotic actions - into a unified token space, modeling them via a single transformer-based denoising network. In this paper, we propose dWorldEval, using a discrete diffusion world model as a scalable evaluation proxy for robotics policy. Specifically, it maps all modalities, including vision, language, and robotics action into a unified token space, then denoises them with a single transformer network. Building on this architecture, we employ a sparse keyframe memory to maintain spatiotemporal consistency. We also introduce a progress token that indicates the degree of task completion. At inference, the model jointly predicts future observations and progress token, allowing automatically determine success when the progress reaches 1. Extensive experiments demonstrate that dWorldEval significantly outperforms previous approaches, i.e., WorldEval, Ctrl-World, and WorldGym, on LIBERO, RoboTwin, and multiple real-robot tasks. It paves the way for a new architectural paradigm in building world simulators for robotics evaluation at scale.

Current benchmarks for graphical user interface (GUI) agents predominantly rely on static screenshots. However, real-world smartphone interaction routinely requires agents to process transient audio cues and temporal video dynamics that are tightly coupled with the moment of action. To bridge this gap, we introduce OmniGUI, the first step-level benchmark designed to evaluate GUI agents in omni-modal smartphone environments. OmniGUI provides continuous, interleaved multimodal inputs comprising static images, synchronous audio, and video clips at every action step. The dataset encompasses 709 expert-demonstrated episodes (2,579 action steps) across 29 applications, systematically annotated with objective multimodal dependency levels. Because dedicated omni-modal GUI agent frameworks are currently in their nascent stage, we select foundational omni-modal models capable of natively processing interleaved inputs to serve as agent proxies for our initial baselines. Our empirical evaluation reveals that while current models exhibit competency on visually static tasks, their action prediction performance degrades significantly in environments requiring synchronous temporal and auditory signals. Furthermore, ablation studies isolate specific operational bottlenecks, notably cross-modal interference when processing task-irrelevant environmental noise. The complete dataset, evaluation pipeline, and baseline prompts are provided in the supplementary material. Project page: https://omni-gui.github.io.

Vision-language-action models (VLAs) have been extensively used in robotics applications, achieving great success in various manipulation problems. More recently, VLAs have been used in long-horizon tasks and evaluated on benchmarks, such as BEHAVIOR1K (B1K), for solving complex household chores. The common metric for measuring progress in such benchmarks is success rate or partial score based on satisfaction of progress-agnostic criteria, meaning only the final states of the objects are considered, regardless of the events that lead to such states. In this paper, we argue that using such evaluation protocols say little about safety aspects of operation and can potentially exaggerate reported performance, undermining core challenges for future real-world deployment. To this end, we conduct a thorough analysis of state-of-the-art models on the B1K Challenge and evaluate policies in terms of robustness via reproducibility and consistency of performance, safety aspects of policies operations, task awareness, and key elements leading to the incompletion of tasks. We then propose evaluation protocols to capture safety violations to better measure the true performance of the policies in more complex and interactive scenarios. At the end, we discuss the limitations of the existing VLAs and motivate future research.

Multimodal Large Language Models (MLLMs) are emerging as core reasoning modules for embodied agents, yet it remains unclear how well general-purpose models can solve long-horizon embodied tasks from a single high-level instruction. We introduce MissionBench, a benchmark for mission-level evaluation of MLLMs in aerial 3D environments. It comprises 120 missions across five simulated 3D environments and four task families. Agents must autonomously plan, navigate, and report outcomes using only egocentric observations and its action history, without aerial-specific fine-tuning. Across 22 open- and closed-source MLLMs, the strongest model succeeds on fewer than 35% of missions compared to 84.4% human performance, highlighting the difficulty of multi-step embodied tasks. Despite large variations between model families, we observe gains from scaling, indicating that larger general-purpose models possess stronger zero-shot embodied capabilities. Our analysis shows that mission-level competence requires coordinating multiple capabilities beyond spatial perception, including multi-step planning and adaptive reasoning. This motivates closed-loop evaluation and highlights both the promise and risk of scaling-driven improvements for embodied AI.

Environments are the foundation of recursive self-improvement: they provide the problems agents work on and the feedback used to evaluate progress. Yet constructing challenging research environments with reliable evaluation still depends on domain experts, limiting their scale and disciplinary coverage. We introduce RSI-Forge, a multi-agent pipeline that turns published papers into executable environments for self-improvement. Three agents coordinate construction, reproduction, and review to produce tasks with automated evaluators; each paper's method is independently reimplemented to establish a baseline score. We present 210 environments across 18 fields, including 90 reviewed by independent human domain experts. Both experts and agent judges give high ratings to the potential for improving the provided starting solutions and the evaluators' ability to distinguish solution quality, whereas experts are more critical of shortcut resistance, faithfulness to the source paper, and whether a single idea can exhaust a task. To validate their use for repeated improvement, we evaluate four models over 3 successive attempts on 120 environments, with each attempt inheriting prior code and notes while model weights remain fixed. At least one model improves after the first attempt in 84% of environments. Models also outperform the reproduced paper methods in 68 of the 120 environments, demonstrating room for gains beyond these baselines. Transcript analysis identifies work beyond parameter tuning in 95% of these successful attempts. Analysis of the resulting trajectories shows that models scoring lower on these tasks explore less, more often accept gains smaller than the reported standard error, and rely more heavily on tuning to the development set. RSI-Forge provides a scalable approach to constructing research environments for training and evaluating self-improving agents.

Evaluating agents by outcomes alone can obscure the capabilities that produce them. This problem is especially pronounced in evolving environments, where outcomes reflect a closed-loop interaction between agent behavior and changing external conditions. We introduce LiveMACEBench, a process-aware benchmark that uses live financial markets as a naturally evolving testbed for persistent LLM agents. Five frontier LLMs operate along continuous trajectories under matched Tool Use, Persistent Memory, Rule Following, and Multi-Agent Collaboration configurations. We evaluate them through both realized outcomes and mechanism-specific diagnostics derived from complete decision traces. Across 30 days of live evaluation, we find a pronounced outcome-capability gap: realized returns often diverge from capability-specific measurements, and similar outcomes can arise from markedly different patterns of mechanism use. Trace-level diagnostics further expose distinct bottlenecks across capabilities, demonstrating that mechanism access, effective mechanism use, and downstream performance are not interchangeable measures of agent capability. LiveMACEBench makes this distinction measurable, turning live markets from a performance leaderboard into a diagnostic environment for agent capability

Lifelong navigation (LN) requires an embodied agent to solve a sequence of navigation subtasks in the same environment. Since solving each subtask from scratch incurs redundant exploration, an LN agent must consolidate experience from earlier stages and reuse it in later stages, often through persistent scene representations such as scene graphs or visual snapshots. However, existing approaches typically assume a stationary environment, whereas in real-world LN settings, human activities can cause the environment to evolve. With the stationary assumption violated, existing methods may fuse outdated prior observations with new observations, yet current benchmarks cannot reveal this failure mode. In this paper, we present EvoNav-Bench, which extends the GOAT-Bench style LN formulation in the context of evolving environments. Built on the ProcTHOR framework, EvoNav-Bench introduces environment modifications between navigation tasks, making prior experience useful but not fully reliable. This design enables controlled evaluation of how environment evolution affects LN agents that reuse prior scene observations. Using EvoNav-Bench, we benchmark three recent methods that build and reuse scene representations for navigation. We also compare three simple heuristic strategies for handling environment evolution: Frontier-Update, Fail-then-Update, and Stage-Reset. Our results show that existing methods are brittle under environment evolution, while the heuristic strategies enable a controlled analysis of how agents can adapt to scene changes and mitigate their impact.

GUI web agents require realistic and interactive environments that remain stable during both evaluation and training. Live website evaluation suffers from changing content, access restrictions, and limited coverage of authenticated workflows, while existing synthetic environments for training often lack functional depth and visual fidelity. We propose WebHarbor, a suite of 153 executable and dockerized website environments that faithfully reconstruct the visual interfaces, content, and functional workflows of real-world web services under deterministic control. We build WebHarbor through task-driven human-agent collaboration, where environments evolve as new tasks expand their functional scope, coding agents implement the required changes, and human contributors audit visual fidelity, functional correctness, and task validity. The resulting suite comprises 3,518 tasks spanning multi-step and post-authentication workflows, where existing works often fail to cover. Beyond reliable evaluation, WebHarbor provides high-fidelity synthetic web environments for cost-efficient reinforcement learning (RL). We demonstrate that a Qwen 4B-level GUI agent trained with RL in WebHarbor generalizes to unseen tasks and performs competitively with leading open-source GUI web agents across four benchmarks. The code and data are publicly available at https://github.com/aiming-lab/WebHarbor.

Diverse evaluation benchmarks play a crucial role to assess a wide range of capabilities of large language models (LLM). Although plenty of endeavors have been dedicated to building valuable benchmarks, there is still little work aiming at evaluating the capability of LLM in multistep interactive environments. Noticing that LLM requires a text representation of the environment observations for interaction, we choose to fill such a blank by building a novel benchmark based on the information user interface (InfoUI). InfoUI consists of rich text contents and can be represented in some text formats, thus is suitable for the assessment of interaction ability of LLM. Additionally, the complex structures of InfoUI can further raise a challenge for LLM to understand structured texts rather than plain texts. An interaction platform is always used to evaluate an agent, however, there is still a lack of a satisfactory interaction platform dedicated to InfoUI. Consequently, we propose to build a novel easily-extendable, adaptable, and close-to-reality interaction platform, Mobile-Env, to provide a base for an appropriate benchmark. Based on Mobile-Env, an InfoUI task set WikiHow is then built to establish a benchmark for the multistep interaction capability of LLM in structured text-based environments. Agents based on a series of LLMs are tested on the task set to obtain an insight into the potential and challenge of LLM for InfoUI interaction. It is sincerely welcome that the community contribute new environments and new task sets for Mobile-Env to provide better test benchmarks and facilitate the development of the corresponding domains.

LiDAR-based localization is valuable for applications like mining surveys and underground facility maintenance. However, existing methods can struggle when dealing with uninformative geometric structures in challenging scenarios. This paper presents RELEAD, a LiDAR-centric solution designed to address scan-matching degradation. Our method enables degeneracy-free point cloud registration by solving constrained ESIKF updates in the front end and incorporates multisensor constraints, even when dealing with outlier measurements, through graph optimization based on Graduated Non-Convexity (GNC). Additionally, we propose a robust Incremental Fixed Lag Smoother (rIFL) for efficient GNC-based optimization. RELEAD has undergone extensive evaluation in degenerate scenarios and has outperformed existing state-of-the-art LiDAR-Inertial odometry and LiDAR-Visual-Inertial odometry methods.

Vision-Language Models (VLMs) have emerged as powerful tools in artificial intelli-gence, capable of integrating textual and visual data for a unified understanding of complex scenes. While models such as Florence2, built on transformer architectures, have shown promise across general tasks, their performance in object detection within unstructured or cluttered environments remains underexplored. In this study, we fi-ne-tuned the Florence2 model for object detection tasks in non-constructed, complex environments. A comprehensive experimental framework was established involving multiple hardware configurations (NVIDIA T4, L4, and A100 GPUs), optimizers (AdamW, SGD), and varied hyperparameters including learning rates and LoRA (Low-Rank Adaptation) setups. Model training and evaluation were conducted on challenging datasets representative of real-world, disordered settings. The optimized Florence2 models exhibited significant improvements in object detection accuracy, with Mean Average Precision (mAP) metrics approaching or matching those of estab-lished models such as YOLOv8, YOLOv9, and YOLOv10. The integration of LoRA and careful fine-tuning of transformer layers contributed notably to these gains. Our find-ings highlight the adaptability of transformer-based VLMs like Florence2 for do-main-specific tasks, particularly in visually complex environments. The study under-scores the potential of fine-tuned VLMs to rival traditional convolution-based detec-tors, offering a flexible and scalable approach for advanced vision applications in re-al-world, unstructured settings.

With the widespread adoption of large language models (LLMs) in code generation tasks, geospatial code generation has emerged as a critical frontier in the integration of artificial intelligence and geoscientific analysis. This trend underscores the urgent need for systematic evaluation methodologies to assess LLMs generation capabilities in geospatial contexts. In particular, geospatial computation and visualization tasks in JavaScript environments rely heavily on orchestrating diverse frontend libraries and ecosystems, placing elevated demands on a model's semantic understanding and code synthesis abilities. To address this challenge, we propose GeoJSEval--the first multimodal, function-level automatic evaluation framework for LLMs in JavaScript-based geospatial code generation. GeoJSEval comprises three core components: a standardized test suite (GeoJSEval-Bench), a code submission engine, and an evaluation module. It includes 432 function-level tasks and 2,071 structured test cases spanning five widely used JavaScript geospatial libraries and 25 mainstream geospatial data types. GeoJSEval enables multidimensional quantitative evaluation across metrics such as accuracy, output stability, execution efficiency, resource consumption, and error type distribution, and integrates boundary testing mechanisms to enhance robustness and coverage. We conduct a comprehensive evaluation of 18 state-of-the-art LLMs using GeoJSEval, revealing significant performance disparities and bottlenecks in spatial semantic understanding, code reliability, and function invocation accuracy. GeoJSEval provides a foundational methodology, evaluation resource, and practical toolkit for the standardized assessment and optimization of geospatial code generation models, with strong extensibility and applicability in real-world scenarios.

The evolution of vehicle-to-everything (V2X) communication brings significant challenges, such as data integrity and vulnerabilities stemming from centralized management. This paper presents an innovative integration of decentralized blockchain technology with V2X communication through a multi-layered architecture that combines the Simulation of Urban Mobility (SUMO) traffic simulator and the BlockSim blockchain simulator. In addition, as the Social Internet of Vehicles (SIoV) emerges, efficient resource management becomes indispensable for ensuring seamless communication. We also propose a reference multi-connectivity management method named Enhanced MAX-SINR, designed to advance research in blockchain-specific approaches, taking into account retransmission successfull rates. We evaluate blockchain performance in diverse environments such as urban, suburban, and rural areas, demonstrating that enhancing the success rate of retransmitted blockchain-related messages significantly boosts blockchain transaction performance and provides a foundation for developing intelligent SIoV systems.

Embodied intelligence is advancing rapidly, driving the need for efficient evaluation. Current benchmarks typically rely on interactive simulated environments or real-world setups, which are costly, fragmented, and hard to scale. To address this, we introduce StaticEmbodiedBench, a plug-and-play benchmark that enables unified evaluation using static scene representations. Covering 42 diverse scenarios and 8 core dimensions, it supports scalable and comprehensive assessment through a simple interface. Furthermore, we evaluate 19 Vision-Language Models (VLMs) and 11 Vision-Language-Action models (VLAs), establishing the first unified static leaderboard for Embodied intelligence. Moreover, we release a subset of 200 samples from our benchmark to accelerate the development of embodied intelligence.

We present WildRayZer, a self-supervised framework for novel view synthesis (NVS) in dynamic environments where both the camera and objects move. Dynamic content breaks the multi-view consistency that static NVS models rely on, leading to ghosting, hallucinated geometry, and unstable pose estimation. WildRayZer addresses this by performing an analysis-by-synthesis test: a camera-only static renderer explains rigid structure, and its residuals reveal transient regions. From these residuals, we construct pseudo motion masks, distill a motion estimator, and use it to mask input tokens and gate loss gradients so supervision focuses on cross-view background completion. To enable large-scale training and evaluation, we curate Dynamic RealEstate10K (D-RE10K), a real-world dataset of 15K casually captured dynamic sequences, and D-RE10K-iPhone, a paired transient and clean benchmark for sparse-view transient-aware NVS. Experiments show that WildRayZer consistently outperforms optimization-based and feed-forward baselines in both transient-region removal and full-frame NVS quality with a single feed-forward pass.

Recent advances in reinforcement learning for code generation have made robust environments essential to prevent reward hacking. As LLMs increasingly serve as evaluators in code-based RL, their ability to detect reward hacking remains understudied. In this paper, we propose a novel taxonomy of reward exploits spanning across 54 categories and introduce TRACE (Testing Reward Anomalies in Code Environments), a synthetically curated and human-verified benchmark containing 517 testing trajectories. Unlike prior work that evaluates reward hack detection in isolated classification scenarios, we contrast these evaluations with a more realistic, contrastive anomaly detection setup on TRACE. Our experiments reveal that models capture reward hacks more effectively in contrastive settings than in isolated classification settings, with GPT-5.2 with highest reasoning mode achieving the best detection rate at 63%, up from 45% in isolated settings on TRACE. Building on this insight, we demonstrate that state-of-the-art models struggle significantly more with semantically contextualized reward hacks compared to syntactically contextualized ones. We further conduct qualitative analyses of model behaviors, as well as ablation studies showing that the ratio of benign to hacked trajectories and analysis cluster sizes substantially impact detection performance. We release the benchmark and evaluation harness to enable the community to expand TRACE and evaluate their models.

This study evaluates the performance of classical and modern control methods for real-world Cable-Driven Parallel Robots (CDPRs), focusing on underconstrained systems with limited time discretization. A comparative analysis is conducted between classical PID controllers and modern reinforcement learning algorithms, including Deep Deterministic Policy Gradient (DDPG), Proximal Policy Optimization (PPO), and Trust Region Policy Optimization (TRPO). The results demonstrate that TRPO outperforms other methods, achieving the lowest root mean square (RMS) errors across various trajectories and exhibiting robustness to larger time intervals between control updates. TRPO's ability to balance exploration and exploitation enables stable control in noisy, real-world environments, reducing reliance on high-frequency sensor feedback and computational demands. These findings highlight TRPO's potential as a robust solution for complex robotic control tasks, with implications for dynamic environments and future applications in sensor fusion or hybrid control strategies.

360 panoramic images are increasingly used in virtual reality, autonomous driving, and robotics for holistic scene understanding. However, current Vision-Language Models (VLMs) struggle with 3D spatial reasoning on Equirectangular Projection (ERP) images due to geometric distortion and limited 3D supervision. We introduce PanoEnv, a large-scale VQA benchmark built from synthetic 3D environments, containing 14.8K questions across five categories (e.g., relative position, volume comparison) grounded in accurate 3D annotations including depth, segmentation, and bounding boxes. Benchmarking 14 state-of-the-art VLMs reveals limited 3D understanding, achieving only 49.34% overall accuracy and 8.36% on open-ended (OE) questions. To enhance 3D reasoning, we propose a reinforcement learning post-training framework based on Group Relative Policy Optimization (GRPO) with a ground-truth-guided reward that incorporates five geometry-aware strategies such as distance tolerance and spatial consistency. A two-stage curriculum further mitigates catastrophic forgetting: Stage 1 trains on structured tasks (true/false and multiple choice), and Stage 2 fine-tunes on mixed open-ended data to improve generalization. Our 7B model achieves new state-of-the-art performance, improving overall accuracy to 52.93% (+3.59%) and open-ended accuracy to 14.83% while maintaining structured-task performance. It also achieves top semantic evaluation scores (Q-Score 6.24, P-Score 5.95), surpassing 32B models. These results demonstrate that PanoEnv-QA and our curriculum-based RL framework effectively instill 3D spatial intelligence in VLMs for omnidirectional perception.

Machine unlearning aims to remove the influence of specific training data from pre-trained models without retraining from scratch, and is increasingly important for large language models (LLMs) due to safety, privacy, and legal concerns. Although prior work primarily evaluates unlearning in static, single-turn settings, forgetting robustness under realistic interactive use remains underexplored. In this paper, we study whether unlearning remains stable in interactive environments by examining two common interaction patterns: self-correction and dialogue-conditioned querying. We find that knowledge appearing forgotten in static evaluation can often be recovered through interaction. Although stronger unlearning improves apparent robustness, it often results in behavioral rigidity rather than genuine knowledge erasure. Our findings suggest that static evaluation may overestimate real-world effectiveness and highlight the need for ensuring stable forgetting under interactive settings.

LLM-based agents increasingly operate in persistent environments where they must store, update, and reason over information across many sessions. While prior benchmarks evaluate only single-entity updates, MEME defines six tasks spanning the full space defined by the multi-entity and evolving axes, including three not scored by prior work: Cascade and Absence (dependency reasoning) and Deletion (post-removal state). Evaluating six memory systems spanning three memory paradigms on 100 controlled episodes, we find that all systems collapse on dependency reasoning under the default configuration (Cascade: 3%, Absence: 1% in average accuracy) despite adequate static retrieval performance. Prompt optimization, deeper retrieval, reduced filler noise, and most stronger LLMs fail to close this gap. Only a file-based agent paired with Claude Opus 4.7 as its internal LLM partially closes the gap, but at ~70x the baseline cost, indicating closure currently depends on configurations that are not practical at scale. Code and data are available on the project page: https://seokwonjung-jay.github.io/meme-eval/.

A central bottleneck for phone-use agents is that controllable, reproducible environments covering real mobile behavior are hard to build at scale. Existing mobile-agent benchmarks have made important progress on evaluation, but they do not by themselves provide a scalable way to construct many new phone-use environments. We present PhoneWorld, a reusable pipeline that converts real GUI trajectories and screenshots into controllable phone-use environments, executable tasks, automatic verifiers, and training rollouts. Rather than hand-building one mobile benchmark at a time, PhoneWorld uses real trajectories to recover which screens matter, how screens connect, which interactions must change environment state, and which user goals admit automatic verification. From these signals, it builds runnable mock Android apps backed by read-only app content and mutable state, then derives executable tasks, rule-based verifiers, and training rollouts from the same environments. In its current instantiation, PhoneWorld covers 34 apps across 16 domains, spanning common consumer mobile behaviors such as search, browsing, shopping, booking, media, and social interaction. Under a fixed training budget, replacing 10K steps from an auxiliary AndroidWorld corpus in an AndroidWorld-based baseline with broad PhoneWorld supervision improves all four evaluation benchmarks at once, raising HYMobileBench by 17.7 points, AndroidControl by 6.0 points, AndroidWorld by 14.7 points, and PhoneWorld by 52.5 points. We then study two additional scaling questions: increasing the amount of PhoneWorld supervision strongly improves PhoneWorld performance, and under a fixed PhoneWorld budget, expanding app coverage yields even larger gains. Overall, PhoneWorld shifts the focus from building one mobile benchmark at a time to scaling the supply of phone-use environments themselves.

In this work, we introduce DeepIPC, a novel end-to-end model tailored for autonomous driving, which seamlessly integrates perception and control tasks. Unlike traditional models that handle these tasks separately, DeepIPC innovatively combines a perception module, which processes RGBD images for semantic segmentation and generates bird's eye view (BEV) mappings, with a controller module that utilizes these insights along with GNSS and angular speed measurements to accurately predict navigational waypoints. This integration allows DeepIPC to efficiently translate complex environmental data into actionable driving commands. Our comprehensive evaluation demonstrates DeepIPC's superior performance in terms of drivability and multi-task efficiency across diverse real-world scenarios, setting a new benchmark for end-to-end autonomous driving systems with a leaner model architecture. The experimental results underscore DeepIPC's potential to significantly enhance autonomous vehicular navigation, promising a step forward in the development of autonomous driving technologies. For further insights and replication, we will make our code and datasets available at https://github.com/oskarnatan/DeepIPC.

Indirect prompt injection poses a critical threat to LLM agents that interact with untrusted external data, yet automated attack methods--proven effective for jailbreaking--remain underexplored in realistic agentic settings. We present a comprehensive empirical evaluation of automated prompt injection attacks against LLM agents, adapting both white-box (GCG) and black-box (TAP) methods to the agentic setting within the AgentDojo framework. We evaluate across 80 task pairs spanning four domains and multiple models, and find that black-box optimization substantially outperforms gradient-based methods, a gap we attribute to GCG's optimization instability under reasonable compute budgets. We also find that TAP's effectiveness depends on the attacker model, as both general capability and safety tuning affect attack success--stronger models produce more effective injections, while safety-tuned attackers can refuse to generate adversarial prompts. Task-universal attacks transfer effectively to unseen tasks and out-of-distribution domains, but attacks optimized on smaller open-source models do not transfer to frontier models like GPT-5. These findings highlight automated prompt injection as a credible but model-dependent threat, with significant barriers remaining for model-agnostic exploitation.

In egocentric action recognition a single population model is typically trained and subsequently embodied on a head-mounted device, such as an augmented reality headset. While this model remains static for new users and environments, we introduce an adaptive paradigm of two phases, where after pretraining a population model, the model adapts on-device and online to the user's experience. This setting is highly challenging due to the change from population to user domain and the distribution shifts in the user's data stream. Coping with the latter in-stream distribution shifts is the focus of continual learning, where progress has been rooted in controlled benchmarks but challenges faced in real-world applications often remain unaddressed. We introduce EgoAdapt, a benchmark for real-world egocentric action recognition that facilitates our two-phased adaptive paradigm, and real-world challenges naturally occur in the egocentric video streams from Ego4d, such as long-tailed action distributions and large-scale classification over 2740 actions. We introduce an evaluation framework that directly exploits the user's data stream with new metrics to measure the adaptation gain over the population model, online generalization, and hindsight performance. In contrast to single-stream evaluation in existing works, our framework proposes a meta-evaluation that aggregates the results from 50 independent user streams. We provide an extensive empirical study for finetuning and experience replay.

As AI agents become increasingly capable of complex, long-horizon reasoning, rigorous and holistic evaluation is essential for measuring progress toward real-world healthcare applications. We introduce HealthAgentBench, a suite of 54 agentic healthcare tasks across 7 categories each with its unique environment. The benchmark suite spans diverse workflows throughout the patient journey and a broad range of modalities. Each task is designed to replicate an end-to-end clinical workflow: given minimal instructions, an agent must explore raw healthcare data, operate within a complex environment, and execute multi-step solutions that go beyond naive prompting. A final task success rate is reported to provide a single, interpretable metric for HealthAgentBench overall performance for each agent. Evaluating frontier agents on HealthAgentBench, we find that overall task success rate remains low, underscoring the difficulty of the suite. The strongest and the most cost effective agent, Codex GPT-5.5, achieves only approximately 42% success rate. Beyond aggregate performance, HealthAgentBench reveals nuanced strengths and weaknesses across task categories. Frontier agents show promise in automatically developing research modeling pipelines over EHR data, but medical imaging remains especially challenging, particularly for Claude Code models, while Codex GPT-5.5 shows emerging capability. Tasks that combine large search spaces with compositional reasoning requirements remain difficult for all current agents. Together, these results suggest that HealthAgentBench provides a challenging and realistic benchmark with substantial room for future progress. We release our benchmark at https://github.com/microsoft/HealthAgentBench.

Collaboration and information sharing empower Multi-Agent Systems (MAS) but also introduce a critical security risk known as Agent Cascading Injection (ACI). In such attacks, a compromised agent exploits inter-agent trust to propagate malicious instructions, causing cascading failures across the system. However, existing studies consider only limited attack strategies and simplified MAS settings, limiting their generalizability and comprehensive evaluation. To bridge this gap, we introduce ACIArena, a unified framework for evaluating the robustness of MAS. ACIArena offers systematic evaluation suites spanning multiple attack surfaces (i.e., external inputs, agent profiles, inter-agent messages) and attack objectives (i.e., instruction hijacking, task disruption, information exfiltration). Specifically, ACIArena establishes a unified specification that jointly supports MAS construction and attack-defense modules. It covers six widely used MAS implementations and provides a benchmark of 1,356 test cases for systematically evaluating MAS robustness. Our benchmarking results show that evaluating MAS robustness solely through topology is insufficient; robust MAS require deliberate role design and controlled interaction patterns. Moreover, defenses developed in simplified environments often fail to transfer to real-world settings; narrowly scoped defenses may even introduce new vulnerabilities. ACIArena aims to provide a solid foundation for advancing deeper exploration of MAS design principles.

Effective agency in social environments depends on when an agent seeks knowledge, when it acts, and whether its actions are justified by acquired information. Existing grounded benchmarks provide executable actions, persistent state, and verifiable outcomes, while social simulation environments provide rich interaction among language agents. We study an evaluation setting that combines these requirements. We define socially distributed task environments as interactive environments where task-relevant knowledge is partitioned across role-isolated participants and consequential actions are accessible only through them. Communication serves as exploration over role-partitioned knowledge, while grounded action serves as exploitation over environment state. We introduce Incognita, a Concordia-based framework that separates social interaction from grounded execution. The evaluated agent routes messages to a user or specialist entities; specialists mediate admissible operations; a deterministic sub-environment executes accepted operations over a canonical state; and an offline evaluator scores outcomes with inherited rewards. Incognita-Retail transforms tau-bench retail into a multi-entity environment while preserving final-state reward semantics. We evaluate three generative agent models on 18 tasks stratified by social breadth, with 540 trials. Progress appears in reward and behavior: success rises from 0 percent to 8.9 percent and 17.2 percent, while premature finalization falls from 100 percent to 87 percent and 58 percent. Stronger models elicit more hidden knowledge, contact more entities, and attempt more grounded writes, yet reliability remains low. These findings show that socially distributed task environments expose behavior before reliable success, including knowledge elicitation, source selection, grounded action attempts, and premature completion belief.

Large language models excel on objectively verifiable tasks such as math and programming, where evaluation reduces to unit tests or a single correct answer. In contrast, real-world enterprise work is often subjective and context-dependent: success hinges on organizational goals, user intent, and the quality of intermediate artifacts produced across long, multi-tool workflows. We introduce LH-Bench, a three-pillar evaluation design that moves beyond binary correctness to score autonomous, long-horizon execution on subjective enterprise tasks. The pillars are: (i) expert-grounded rubrics that give LLM judges the domain context needed to score subjective work, (ii) curated ground-truth artifacts that enable stepwise reward signals (e.g., chapter-level annotation for content tasks), and (iii) pairwise human preference evaluation for convergent validation. We show that domain-authored rubrics provide substantially more reliable evaluation signals than LLM-authored rubrics (kappa = 0.60 vs. 0.46), and that human preference judgments confirm the same top-tier separation (p < 0.05), evidence that expert-grounded evaluation can scale without sacrificing reliability. We release public datasets and report results on two environments: Figma-to-code (33 real .fig tasks against the Figma API via MCP) and Programmatic content (41 courses comprising 183 individually-evaluated chapters on a course platform serving 30+ daily users).

Agents learn to act through interaction with environments, yet the environments used for training are often manually constructed or synthesized around predefined tasks and benchmarks. This task-centric paradigm makes it difficult to scale environments that reflect realistic and evolving workflows where diverse tasks can naturally emerge from the underlying world. We introduce AgentMercury, a scalable framework for synthesizing executable environments from high-level business scenarios. Rather than constructing an environment for a specific task, AgentMercury first instantiates a persistent world with entities, services, tools, state, and executable cross-service invariants, from which diverse tasks and interaction trajectories can subsequently emerge. We construct 4,783 executable environments spanning 14 industries and 50 countries, and use them as training substrates for reinforcement learning. Despite being generated without targeting the evaluation benchmarks, policies trained on these business-oriented environments improve substantially on both enterprise workflows and out-of-domain benchmarks spanning reasoning, coding, scientific computing, and tool use. In our experiments, Qwen3.5-4B improves from 12.3 to 15.7 on EnterpriseOps-GYM and from 45.9 to 56.0 on AIME26 after training on AgentMercury environments. We further show that the construction process itself can be learned: fine-tuning Qwen3.5-35B-A3B on construction traces increases executable-world authoring success from 3.3% to 83.3% on held-out business scenarios. These results show that scenario-grounded environments can provide useful and generalizable learning signals beyond benchmark-specific training, while their construction can itself become a learnable capability.

This paper presents a novel data-driven approach to camera-based autonomy for micro-drones in GPS-denied, radio-challenging indoor environments. The target application is disaster and emergency response, where micro-UAVs can provide rapid situational awareness in hazardous settings such as firefighting and chemical, biological, radiological, and nuclear (CBRN) incidents while reducing risk for human responders. When the communication link is lost, the micro-drone uses a learned yaw controller to autonomously navigate toward open space, preserving onboard sensor data that would otherwise be lost with the vehicle. A custom micro-drone equipped with a 360-degree camera was used to record diverse industrial, underground, and training scenarios representative of communication-denied field operations. We introduce a preprocessing pipeline that converts equirectangular 360-degree footage into planar front views and dynamically generates image-label pairs for AI training. We then train and compare multiple convolutional neural network variants that predict a continuous yaw command from a single monocular view. Evaluation on a held-out test set confirms the feasibility of the learned yaw-prediction approach. A semi-autonomous real-world test further demonstrates the practicality of the method while revealing key failure modes, particularly reflections and glare.

Open-vocabulary visual grounding enables robots to localise task-relevant entities from natural-language queries without dependence on predefined perceptual taxonomies. However, existing benchmarks largely rely on short category labels and web-scraped imagery, leaving it unclear whether open-vocabulary models can robustly ground diverse queries and visual conditions under real deployments. We introduce Pro-Bench, a prompt-conditioned benchmark for open-vocabulary visual grounding in heterogeneous, real-world environments. Pro-Bench includes 13k+ RGB frames from independent robotic domains (subterranean, industrial, indoor, outdoor, urban), with 74.5k manual instance annotations and 515 target queries covering categorical, attributive, relational, affordance, state, part-whole, negative, and compositional semantics. We benchmarked 16 open-vocabulary model configurations in strict zero-shot inference, measuring localisation accuracy across IoU thresholds, end-to-end inference latency, prompt-induced performance variation, and target recovery consistency. Our results show that prompt-robustness is strongly architecture-dependent. Most model configurations (10/16) perform best with short category labels, whereas free-form queries yield the highest accuracy for only one. Moreover, similar aggregate mAP can conceal substantial differences in consistent target recovery across reformulations. Pro-Bench enables systematic evaluation of these gaps and supports prompt-robust visual grounding. Pro-Bench: https://pro-bench.github.io/.

Simulated evaluation is increasingly used alongside real-world evaluation of robot policies because it is cheaper and easier to repeat; however, its value depends on how closely its outcomes track the real robot's. We test whether our reconstruction pipeline, combining metrically scaled object geometry, authored physical parameters and scene reconstruction, reduces disagreement between simulated and real robot scores relative to a default open-source recipe. We constructed two simulated versions of one bimanual robot cell: an authored reconstruction, using object geometry at estimated metric scale, projected textures, authored physics and our own scene splat; and a baseline, referred to as the default reconstruction, using the open-source recipe of a generative single-image mesh, engine-default physics and a Gaussian-splat scene. Both reconstructions use the same object photographs and scene video. Two policies ran five tasks each, giving ten task-policy pairs, which we call cells; each cell was run twenty times in each reconstruction with all other settings held fixed. Both reconstructions were scored against the same real trials, graded by a third-party evaluator. Pearson correlation between the ten simulated and real cell means is r = 0.90 for the authored reconstruction and 0.51 for the default. Mean score error is 6.97 percentage points for the authored reconstruction and 17.54 for the default, a reduction of 10.56 percentage points. These results show that improving the quality of the environment reconstruction through higher visual fidelity, authored physics and metric scale makes the simulation more faithful to the real world and narrows the sim-to-real gap. We release the harness, the per-trial scores, every reported run's configuration, and the assets and scenes of both reconstructions.

As LLM-based agents are deployed with increasing autonomy in real-world environments, a critical failure mode emerges: agentic misalignment, where agents choose harmful actions over failure. However, existing safety benchmarks primarily evaluate whether agents comply with explicitly harmful instructions, leaving the conditions that cause agents to choose harmful actions poorly understood. To address this limitation, we propose AM-Bench, a unified taxonomy and 35 scenario evaluation benchmark spanning seven industries for agentic misalignment grounded in the fraud triangle, a well-established criminological framework explaining goal-directed misconduct. We adapt this framework to agentic environments and construct our evaluation suite and detection systems along three dimensions: perceived threat, perceived opportunity, and reasoning. We operationalize each dimension into structured subcategories rooted in prior theory: five threat categories derived from the instrumental convergence thesis, seven opportunity categories drawn from the Emergent Strategic Reasoning Risks (ESRR) taxonomy, and a novel four-question binary reasoning framework synthesized from prior literature. Our evaluation of 10 frontier models reveals a 62-percentage-point gap in misalignment across scenarios, with Gemini 3 Flash and Gemini 3.1 Pro exhibiting the highest levels of agentic misalignment while Claude Sonnet 4.6 displays near-zero rates across all conditions. Our reasoning framework further reveals that models may exhibit harmful behaviours without explicit situational awareness or explanation, complicating both mitigation and detection. AM-Bench scenarios and code are available at https://github.com/AI4Collaboration/Agentic-Misalignment-Bench

Despite substantial progress in Deep Search agents, their ability to conduct end-to-end research in real-world open Web environments remains insufficiently evaluated. Unlike static retrieval settings, information on the real-world Web is distributed across diverse sources, modalities, and heterogeneous formats, often involving complex interaction mechanisms that require agents to autonomously search for, access, and integrate relevant evidence during execution. To address this gap, we introduce the OpenDS task, designed to evaluate agents' ability to perform multi-step deep search in the live, open Web. This task requires agents to autonomously explore the real Web, discover and connect relevant evidence from multiple independent sources through Web search, browser interaction, and evidence processing, and ultimately determine a unique factual answer. Building on this task formulation, we construct OpenDS-World, a benchmark comprising 200 challenging tasks spanning diverse information modalities, heterogeneous file formats, and interactive access requirements, with independent human verification to ensure evidence validity and answer uniqueness. We evaluate 12 frontier models under a unified tool environment and execution protocol. The best-performing model achieves 39.0% accuracy, while human evaluation across 1,000 independent attempts yields 21.1% accuracy. Further analyses of tool usage, search cost, termination behavior, and failure recovery reveal substantial differences in research strategies across models, as well as persistent challenges in tool coordination and long-horizon search management. OpenDS-World provides a unified evaluation platform for systematically assessing and advancing agents' autonomous deep search capabilities in real-world open Web environments.

Topological mapping offers a compact and robust representation for navigation, but progress in the field is hindered by the lack of standardized evaluation metrics, datasets, and protocols. Existing systems are assessed using different environments and criteria, preventing fair and reproducible comparisons. Moreover, a key challenge - perceptual aliasing - remains under-quantified, despite its strong influence on system performance. We address these gaps by (1) formalizing topological consistency as the fundamental property of topological maps and showing that localization accuracy provides an efficient and interpretable surrogate metric, and (2) proposing the first quantitative measure of dataset ambiguity to enable fair comparisons across environments. To support this protocol, we curate a diverse benchmark dataset with calibrated ambiguity levels, implement and release deep-learned baseline systems, and evaluate them alongside classical methods. Our experiments and analysis yield new insights into the limitations of current approaches under perceptual aliasing. All datasets, baselines, and evaluation tools are fully open-sourced to foster consistent and reproducible research in topological mapping.

Neural networks often suffer from catastrophic interference (CI): performance on previously learned tasks drops off significantly when learning a new task. This contrasts strongly with humans, who can sequentially learn new tasks without appreciably forgetting previous tasks. Prior work has explored various techniques for mitigating CI such as regularization, rehearsal, generative replay, and distillation methods. The current work takes a different approach, one guided by cognitive science research showing that in naturalistic environments, the probability of encountering a task decreases as a power-law of the time since it was last performed. We argue that a realistic evaluation of techniques for the mitigation of CI should be performed in simulated naturalistic learning environments. Thus, we evaluate the extent of mitigation of CI when training simple rehearsal-based methods in power-law environments similar to the ones humans face. Our work explores this novel rehearsal-based approach for a domain-incremental task: learning permutations in the MNIST task. We compare our rehearsal environment with other baselines to show its efficacy in promoting continual learning. Additionally, we investigate whether this environment shows forward facilitation, i.e., faster learning of later tasks. Next, we explore the robustness of our learning environment to the number of tasks, model size, and amount of data rehearsed after each task. Notably, our results show that the performance is comparable or superior to that of models trained using popular regularization methods and also to rehearsals in non-power-law environments. The benefits of this training paradigm include simplicity and the lack of a need for extra neural circuitry. In addition, because our method is orthogonal to other methods, future research can combine training in power-law environments with other continual learning mechanisms.

Code completion has become an essential tool for daily software development. Existing evaluation benchmarks often employ static methods that do not fully capture the dynamic nature of real-world coding environments and face significant challenges, including limited context length, reliance on superficial evaluation metrics, and potential overfitting to training datasets. In this work, we introduce a novel framework for enhancing code completion in software development through the creation of a repository-level benchmark ExecRepoBench and the instruction corpora Repo-Instruct, aim at improving the functionality of open-source large language models (LLMs) in real-world coding scenarios that involve complex interdependencies across multiple files. ExecRepoBench includes 1.2K samples from active Python repositories. Plus, we present a multi-level grammar-based completion methodology conditioned on the abstract syntax tree to mask code fragments at various logical units (e.g. statements, expressions, and functions). Then, we fine-tune the open-source LLM with 7B parameters on Repo-Instruct to produce a strong code completion baseline model Qwen2.5-Coder-Instruct-C based on the open-source model. Qwen2.5-Coder-Instruct-C is rigorously evaluated against existing benchmarks, including MultiPL-E and ExecRepoBench, which consistently outperforms prior baselines across all programming languages. The deployment of can be used as a high-performance, local service for programming development\url{https://execrepobench.github.io/}.

Progress in machine learning is measured by careful evaluation on problems of outstanding common interest. However, the proliferation of benchmark suites and environments, adversarial attacks, and other complications has diluted the basic evaluation model by overwhelming researchers with choices. Deliberate or accidental cherry picking is increasingly likely, and designing well-balanced evaluation suites requires increasing effort. In this paper we take a step back and propose Nash averaging. The approach builds on a detailed analysis of the algebraic structure of evaluation in two basic scenarios: agent-vs-agent and agent-vs-task. The key strength of Nash averaging is that it automatically adapts to redundancies in evaluation data, so that results are not biased by the incorporation of easy tasks or weak agents. Nash averaging thus encourages maximally inclusive evaluation -- since there is no harm (computational cost aside) from including all available tasks and agents.

Recently, the DeepSeek series of models, leveraging their exceptional reasoning capabilities and open-source strategy, is reshaping the global AI landscape. Despite these advantages, they exhibit significant safety deficiencies. Research conducted by Robust Intelligence, a subsidiary of Cisco, in collaboration with the University of Pennsylvania, revealed that DeepSeek-R1 has a 100\% attack success rate when processing harmful prompts. Additionally, multiple safety companies and research institutions have confirmed critical safety vulnerabilities in this model. As models demonstrating robust performance in Chinese and English, DeepSeek models require equally crucial safety assessments in both language contexts. However, current research has predominantly focused on safety evaluations in English environments, leaving a gap in comprehensive assessments of their safety performance in Chinese contexts. In response to this gap, this study introduces CHiSafetyBench, a Chinese-specific safety evaluation benchmark. This benchmark systematically evaluates the safety of DeepSeek-R1 and DeepSeek-V3 in Chinese contexts, revealing their performance across safety categories. The experimental results quantify the deficiencies of these two models in Chinese contexts, providing key insights for subsequent improvements. It should be noted that, despite our efforts to establish a comprehensive, objective, and authoritative evaluation benchmark, the selection of test samples, characteristics of data distribution, and the setting of evaluation criteria may inevitably introduce certain biases into the evaluation results. We will continuously optimize the evaluation benchmark and periodically update this report to provide more comprehensive and accurate assessment outcomes. Please refer to the latest version of the paper for the most recent evaluation results and conclusions.

Recent advancements in Vision Language Models (VLMs) have expanded their capabilities to interactive agent tasks, yet existing benchmarks remain limited to single-agent or text-only environments. In contrast, real-world scenarios often involve multiple agents interacting within rich visual and linguistic contexts, posing challenges with both multimodal observations and strategic interactions. To bridge this gap, we introduce Visual Strategic Bench (VS-Bench), a multimodal benchmark that evaluates VLMs for strategic reasoning and decision-making in multi-agent environments. VS-Bench comprises eight vision-grounded environments spanning cooperative, competitive, and mixed-motive interactions, designed to assess agents' ability to predict others' future moves and optimize for long-term objectives. We consider two complementary evaluation dimensions, including offline evaluation of strategic reasoning by next-action prediction accuracy and online evaluation of decision-making by normalized episode return. Extensive experiments of fourteen leading VLMs reveal a significant gap between current models and optimal performance, with the best models attaining 47.8% prediction accuracy and 24.3% normalized return. We further conduct in-depth analyses on multimodal observations, test-time scaling, social behaviors, and failure cases of VLM agents. By standardizing the evaluation and highlighting the limitations of existing models, we envision VS-Bench as a foundation for future research on strategic multimodal agents. Code and data are available at https://vs-bench.github.io.

The deep learning models has significantly advanced dexterous manipulation techniques for multi-fingered hand grasping. However, the contact information-guided grasping in cluttered environments remains largely underexplored. To address this gap, we have developed a method for generating multi-fingered hand grasp samples in cluttered settings through contact semantic map. We introduce a contact semantic conditional variational autoencoder network (CoSe-CVAE) for creating comprehensive contact semantic map from object point cloud. We utilize grasp detection method to estimate hand grasp poses from the contact semantic map. Finally, an unified grasp evaluation model PointNetGPD++ is designed to assess grasp quality and collision probability, substantially improving the reliability of identifying optimal grasps in cluttered scenarios. Our grasp generation method has demonstrated remarkable success, outperforming state-of-the-art methods by at least 4.65% with 81.0% average grasping success rate in real-world single-object environment and 75.3% grasping success rate in cluttered scenes. We also proposed the multi-modal multi-fingered grasping dataset generation method. Our multi-fingered hand grasping dataset outperforms previous datasets in scene diversity, modality diversity. The dataset, code and supplementary materials can be found at https://sites.google.com/view/contact-dexnet.