VibeWorlding: Can Multimodal Agents Construct 3D Open Worlds End-to-End?

Constructing an interactive 3D open world from a user query is important. However, existing methods are primarily evaluated on idealized, simple queries, making it difficult to systematically analyze and compare how multimodal agents understand user intent, use 3D tools, and reason over textual and visual 3D world information. To this end, we propose VibeWorlding, a unified framework for benchmarking and training vibe worlding agents: a multimodal agent that can autonomously infer user intent, plan scene layout, invoke 3D tools, and reflect on the multimodal feedback in a multi-turn agent-environment interaction process. To achieve this, we first build VWE-BENCH, a benchmark of 2,616 high-quality 3D assets, 323 human-annotated seed 3D worlds, and 6,828 reverse-synthesized multimodal user queries, split into verified queries with ground-truth and unverified queries with carefully designed rubrics. Moreover, we develop VibeWorlding-Gym, a joint multimodal RL post-training framework that integrates (1) a sandbox environment unifying asset retrieval, editing, and image rendering as MCP tools, and (2) a rubric-based verifier that combines physical feasibility and intent fulfillment verification, supporting both fair model evaluation and scalable multimodal RL reward service. Our experiments show that current frontier MLLMs are far from solving the vibe worlding agent task, with even GPT-5.5 and Qwen3.8-Max reaching below 60% success rate, and trace the bottleneck to precise 3D world editing. We further find that RL training can ease this weakness and enable open-source MLLMs to even surpass closed-source frontiers: our VibeWorlder-8B is comparable to frontier MLLMs, while our flagship VibeWorlder-30B-A3B attains the best overall Pass@1 among all evaluated models.

EvoDS: Self-Evolving Autonomous Data Science Agent with Skill Learning and Context Management

Recent progress in Large Language Model (LLM) agents has enabled promising advances in automated data science. However, existing approaches remain fundamentally limited by their static action sets and lack of principled long-horizon context management, hindering their ability to accumulate reusable experience across tasks and operate reliably in multi-stage, iterative data science pipelines. To address these challenges, we introduce EvoDS, a self-evolving autonomous data science agent that learns to expand its skills and adaptively managing long-term context through agentic reinforcement learning. Specifically, EvoDS introduces two key strategies: (1) Autonomous Skill Acquisition (ASA) mechanism, which enables agents to synthesize, validate, and reuse executable skills; and (2) Adaptive Context Compression (ACC) strategy, which treats context management as a learned control problem rather than passive truncation. These strategies are orchestrated within a two-stage multi-agent training scheme, enabling EvoDS to autonomously improve over time. Theoretically, we prove that EvoDS's hierarchical design reduces tool-selection error, and its optimization objective aligns with an information bottleneck principle, ensuring efficient context use. Empirically, EvoDS outperforms state-of-the-art open-source data science agents by an average of 28.9% across four diverse benchmarks while eliminating out-of-token failures. Our code and data are available at https://github.com/usail-hkust/EvoDS.

HRBench: Benchmarking and Understanding Thinking-Mode Switch Strategies in Hybrid-Reasoning LLMs

Hybrid-reasoning large language models (LLMs) expose explicit controls over reasoning effort, allowing users or systems to trade off answer quality against inference cost. However, existing methods for adaptive thinking-mode selection are typically evaluated under different models, datasets, and implementation assumptions, making it difficult to compare their practical behavior. We introduce HRBench, a unified evaluation framework for studying thinking-mode switching in hybrid-reasoning LLMs. HRBench organizes the design space along two axes: three switching strategy families, prompt-based selection, external routing, and speculative execution, and four training regimes, training-free, SFT, offline and online RL, yielding 12 controlled evaluation settings. We evaluate these settings across 6 LLMs, from Qwen3.5-2B to Kimi-K2.5-1.1T, and 5 reasoning benchmarks covering mathematics, science, and code, while reimplementing 12+ representative prior methods within the same pipeline. Our analysis characterizes how different switching strategies occupy distinct effectiveness-efficiency trade-off regions: prompt-based methods often provide favorable token-accuracy trade-offs, routing methods offer more stable cost reduction, and speculative methods tend to improve accuracy at higher token cost. We further find that training affects strategies differently, and that the preferred strategy varies with model scale and task domain. HRBench provides reference implementations and a unified evaluation platform to support more controlled research on efficient reasoning in hybrid-reasoning LLMs. Our data, code and repository are available at https://github.com/usail-hkust/HRBench.

TrafficClaw: Generalizable Urban Traffic Control via Unified Physical Environment Modeling

Urban traffic control is a system-level coordination problem spanning heterogeneous subsystems, including traffic signals, freeways, public transit, and taxi services. Existing optimization-based, reinforcement learning (RL), and emerging LLM-based approaches are largely designed for isolated tasks, limiting both cross-task generalization and the ability to capture coupled physical dynamics across subsystems. We argue that effective system-level control requires a unified physical environment in which subsystems share infrastructure, mobility demand, and spatiotemporal constraints, allowing local interventions to propagate through the network. To this end, we propose TrafficClaw, a framework for general urban traffic control built upon a unified runtime environment. TrafficClaw integrates heterogeneous subsystems into a shared dynamical system, enabling explicit modeling of cross-subsystem interactions and closed-loop agent-environment feedback. Within this environment, we develop an LLM agent with executable spatiotemporal reasoning and reusable procedural memory, supporting unified diagnostics across subsystems and continual strategy refinement. Furthermore, we introduce a multi-stage training pipeline with supervised initialization and agentic RL with system-level optimization, further enabling coordinated and system-aware performance. Experiments demonstrate that TrafficClaw achieves robust, transferable, and system-aware performance across unseen traffic scenarios, dynamics, and task configurations. Our project is available at https://github.com/usail-hkust/TrafficClaw.

Agent-Omit: Training Efficient LLM Agents for Adaptive Thought and Observation Omission via Agentic Reinforcement Learning

Managing agent thought and observation during multi-turn agent-environment interactions is an emerging strategy to improve agent efficiency. However, existing studies treat the entire interaction trajectories equally, overlooking the thought necessity and observation utility varies across turns. To this end, we first conduct quantitative investigations into how thought and observation affect agent effectiveness and efficiency. Based on our findings, we propose Agent-Omit, a unified training framework that empowers LLM agents to adaptively omit redundant thoughts and observations. Specifically, we first synthesize a small amount of cold-start data, including both single-turn and multi-turn omission scenarios, to fine-tune the agent for omission behaviors. Furthermore, we introduce an omit-aware agentic reinforcement learning approach, incorporating a dual sampling mechanism and a tailored omission reward to incentivize the agent's adaptive omission capability. Theoretically, we prove that the deviation of our omission policy is upper-bounded by KL-divergence. Experimental results on five agent benchmarks show that our constructed Agent-Omit-8B could obtain performance comparable to seven frontier LLM agent, and achieve the best effectiveness-efficiency trade-off than seven efficient LLM agents methods. Our code and data are available at https://github.com/usail-hkust/Agent-Omit.

AgentBalance: Backbone-then-Topology Design for Cost-Effective Multi-Agent Systems under Budget Constraints

Large Language Model (LLM)-based multi-agent systems (MAS) are becoming indispensable building blocks for web-scale applications such as web search, social network analytics, and online customer support, where cost-effectiveness is increasingly the primary constraint for large-scale deployment. While recent work improves MAS cost-effectiveness by shaping inter-agent communication topologies and selecting agent backbones, it rarely models and optimizes under explicit token-cost and latency budgets that reflect deployment constraints. This often leads to topology-first designs and suboptimal cost-effectiveness when budgets are binding. We present AgentBalance, a framework for constructing cost-effective MAS under explicit token-cost and latency budgets via a backbone-then-topology design. AgentBalance first performs backbone-oriented agent generation, constructing agents with heterogeneous backbones through LLM pool construction, pool selection, and role-backbone matching. It then performs adaptive MAS topology generation, guiding inter-agent communication via agent representation learning, gating, and latency-aware topology synthesis. Experiments on benchmarks with 14 candidate LLM backbones show that AgentBalance achieves up to 10% and 22% performance gains under matched token-cost and latency budgets, respectively, and yields strong AUC on performance-versus-budget curves across benchmarks. AgentBalance also functions as a plug-in for existing MAS, improving performance under the same token-cost and latency constraints, and it generalizes well to unseen LLMs for practical, budget-aware deployment. Code: https://github.com/usail-hkust/AgentBalance

Large Language Model Powered Intelligent Urban Agents: Concepts, Capabilities, and Applications

The long-standing vision of intelligent cities is to create efficient, livable, and sustainable urban environments using big data and artificial intelligence technologies. Recently, the advent of Large Language Models (LLMs) has opened new ways toward realizing this vision. With powerful semantic understanding and reasoning capabilities, LLMs can be deployed as intelligent agents capable of autonomously solving complex problems across domains. In this article, we focus on Urban LLM Agents, which are LLM-powered agents that are semi-embodied within the hybrid cyber-physical-social space of cities and used for system-level urban decision-making. First, we introduce the concept of urban LLM agents, discussing their unique capabilities and features. Second, we survey the current research landscape from the perspective of agent workflows, encompassing urban sensing, memory management, reasoning, execution, and learning. Third, we categorize the application domains of urban LLM agents into five groups: urban planning, transportation, environment, public safety, and urban society, presenting representative works in each group. Finally, we discuss trustworthiness and evaluation issues that are critical for real-world deployment, and identify several open problems for future research. This survey aims to establish a foundation for the emerging field of urban LLM agents and to provide a roadmap for advancing the intersection of LLMs and urban intelligence. A curated list of relevant papers and open-source resources is maintained and continuously updated at https://github.com/usail-hkust/Awesome-Urban-LLM-Agents.

USTBench: Benchmarking and Dissecting Spatiotemporal Reasoning of LLMs as Urban Agents

Large language models (LLMs) have shown emerging potential in spatiotemporal reasoning, making them promising candidates for building urban agents that support diverse urban downstream applications. Despite these benefits, existing studies primarily focus on evaluating urban LLM agent on outcome-level metrics (e.g., prediction accuracy, traffic efficiency), offering limited insight into their underlying reasoning processes. As a result, the strengths and limitations of urban LLM agents in spatiotemporal reasoning remain poorly understood. To this end, we introduce USTBench, the first benchmark to evaluate LLMs' spatiotemporal reasoning abilities as urban agents across four decomposed dimensions: spatiotemporal understanding, forecasting, planning, and reflection with feedback. Specifically, USTBench supports five diverse urban decision-making and four spatiotemporal prediction tasks, all running within our constructed interactive city environment UAgentEnv. The benchmark includes 62,466 structured QA pairs for process-level evaluation and standardized end-to-end task assessments, enabling fine-grained diagnostics and broad task-level comparison across diverse urban scenarios. Through extensive evaluation of thirteen leading LLMs, we reveal that although LLMs show promising potential across various urban downstream tasks, they still struggle in long-horizon planning and reflective adaptation in dynamic urban contexts. Notably, recent advanced reasoning models (e.g., DeepSeek-R1) trained on general logic or mathematical problems do not consistently outperform non-reasoning LLMs. This discrepancy highlights the need for domain-specialized adaptation methods to enhance urban spatiotemporal reasoning. Overall, USTBench provides a foundation to build more adaptive and effective LLM-based urban agents and broad smart city applications.

Not All Thoughts are Generated Equal: Efficient LLM Reasoning via Multi-Turn Reinforcement Learning

Compressing long chain-of-thought (CoT) from large language models (LLMs) is an emerging strategy to improve the reasoning efficiency of LLMs. Despite its promising benefits, existing studies equally compress all thoughts within a long CoT, hindering more concise and effective reasoning. To this end, we first investigate the importance of different thoughts by examining their effectiveness and efficiency in contributing to reasoning through automatic long CoT chunking and Monte Carlo rollouts. Building upon the insights, we propose a theoretically bounded metric to jointly measure the effectiveness and efficiency of different thoughts. We then propose Long$\otimes$Short, an efficient reasoning framework that enables two LLMs to collaboratively solve the problem: a long-thought LLM for more effectively generating important thoughts, while a short-thought LLM for efficiently generating remaining thoughts. Specifically, we begin by synthesizing a small amount of cold-start data to fine-tune LLMs for long-thought and short-thought reasoning styles, respectively. Furthermore, we propose a synergizing-oriented multi-turn reinforcement learning, focusing on the model self-evolution and collaboration between long-thought and short-thought LLMs. Experimental results show that our method enables Qwen2.5-7B and Llama3.1-8B to achieve comparable performance compared to DeepSeek-R1-Distill-Qwen-7B and DeepSeek-R1-Distill-Llama-8B, while reducing token length by over 80% across the MATH500, AIME24/25, AMC23, and GPQA Diamond benchmarks. Our data and code are available at https://github.com/yasNing/Long-otimes-Short/.

UrbanKGent: A Unified Large Language Model Agent Framework for Urban Knowledge Graph Construction

Urban knowledge graph has recently worked as an emerging building block to distill critical knowledge from multi-sourced urban data for diverse urban application scenarios. Despite its promising benefits, urban knowledge graph construction (UrbanKGC) still heavily relies on manual effort, hindering its potential advancement. This paper presents UrbanKGent, a unified large language model agent framework, for urban knowledge graph construction. Specifically, we first construct the knowledgeable instruction set for UrbanKGC tasks (such as relational triplet extraction and knowledge graph completion) via heterogeneity-aware and geospatial-infused instruction generation. Moreover, we propose a tool-augmented iterative trajectory refinement module to enhance and refine the trajectories distilled from GPT-4. Through hybrid instruction fine-tuning with augmented trajectories on Llama-2-13B, we obtain the UrbanKGC agent, UrbanKGent-13B. We perform a comprehensive evaluation on two real-world datasets using both human and GPT-4 self-evaluation. The experimental results demonstrate that UrbanKGent-13B not only can significantly outperform 21 baselines in UrbanKGC tasks, but also surpass the state-of-the-art LLM, GPT-4, by more than 10\% with approximately 20 times lower cost. We deploy UrbanKGent-13B to provide online services, which can construct an UrbanKG with thousands of times richer relationships using only one-fifth of the data compared with the existing benchmark. Our data, code, and opensource UrbanKGC agent are available at https://github.com/usail-hkust/UrbanKGent.