Agent 进展

Although large language models (LLMs) have demonstrated remarkable capabilities, their reliance on cloud-scale infrastructure poses fundamental challenges for deployment in agentic pipelines, including latency, privacy, connectivity, and substantial computational cost. Small language models (SLMs) offer a compelling alternative: recent studies suggest that many repetitive and narrowly scoped subtasks in agentic workloads may be better served by specialized SLMs than by monolithic LLMs. However, the limited capacity and context windows of SLMs can constrain long-horizon reasoning and interaction-heavy orchestration strategies such as iterative verification and debate. This motivates a complementary, non-interactive paradigm in which heterogeneous SLMs independently generate candidate solutions and a router orchestrates their cached samples without further model interaction. To further understand the mechanisms of such orchestration, we introduce OrchSLM, a routing framework that unifies existing non-interactive orchestration methods and exposes their underlying design choices as controllable parameters. Using OrchSLM as a systematic probe, we reveal how orchestration behavior emerges from diverse knobs, including the task structure, model-pool composition, and multi-agent consensus.

Affective computing has advanced wearable state inference, but on-device reasoning about whether, when, and how to intervene remains challenging. We present Affective Agent, a three-layer reference architecture for personalized intervention reasoning under uncertainty on wearable-class hardware. It combines a compact sub-billion-parameter language model with physiological evidence, context, and user history to decide whether, when, and how to intervene, without cloud dependency or per-user retraining. The architecture is organized into three interacting layers (perception, personalization, and reasoning), adapting to individual users through host-managed structured memory evolution rather than per-user weight updates. We instantiate Affective Agent in indoor environmental quality control and evaluate it on held-out, simulator-generated longitudinal scenarios spanning physiological variation, context, signal quality, and intervention history. Results show that memory-driven personalization and two-pass structured reasoning improve intervention decisions within this synthetic evaluation. By moving the decision layer on-device, this work demonstrates a path from wearable state inference toward closed-loop, personalized intervention on wearable-class hardware.

We aim to characterise the value of artificial intelligence in the workplace. Current studies largely measure this value in terms of the current automation capabilities and public adoption of AI. However, such metrics ignore the greater impacts of human--agent collaboration in transforming the nature of work. To account for this, we must expand the scope of our analysis beyond atomised tasks of today, and instead focus on how AI can augment entire workflows of the future. To ground this analysis, we establish a precise definition of AI augmentation comprising six conditions, spanning durable net value, meaningful human control, accountability and recovery, and long-term human development through learning, career pathways, and job purpose. We elaborate on these conditions and apply the framework in a case study of AI-mediated social surveys. We conclude by outlining how organisations, researchers, and government leaders can use this framework to make sense of the future of work.

When an autonomous AI agent does something consequential, what can be proven about what it did? Agent-observability platforms capture traces, but a trace is mutable: alterable undetected, with no recipe for re-executing it, silent on whether captured secrets were removed. Regulation (EU AI Act, ISO 42001, NIST AI RMF) presumes records an independent party can check. We present NovaFabric, producing audit-grade execution evidence: provider-neutral, tamper-evident, replayable, shareable. It records an agent run, without modifying agent logic, into a portable Run Capsule (fifteen-entity schema), sealed with a holistic DSSE signature, RFC 3161 timestamp, Merkle log and redaction attestation. Sealed runs are re-executable under a four-mode replay protocol and exportable as an Evidence Bundle for third-party verification with stock tooling (specified, not evaluated). The contribution is integration, not new cryptography: OpenTelemetry, DSSE/in-toto and W3C PROV. We evaluate eight research questions at measured scope. Mocked replay serves every model response from the capsule (no live model call, 10/10) but is offline w.r.t. models, not the network; only 2/10 tool-using workloads completed; the gap is missing tool-response substitution. Tampering is rejected across three tested classes. Declared-stream completeness is 0.652 (95% CI +/-0.064, ten scenarios). A repaired rule pack redacts 14/14 credential types, preserving 9/9 decoys; diff localises 140/140 mutations. Blast-radius queries: 45.5ms p99 over 10M edges (3.3x faster than a columnar baseline), 167.9ms over 100M (1 client, n=30). A 314-machine, ten-region run finds capsule REST ingest lossless but capped at 61.6 req/s (p99 26.8s) by per-worker serialisation. Six defects found in NovaFabric and its evaluation corpus: four fixed, one withdrawn, one open. Verification is conditional on a stated trusted computing base.

Dementia affects an estimated 57 million people worldwide, and for most families the hardest part of care is not memory loss but the behavioral and psychological symptoms of dementia (BPSD): agitation, wandering, resistance to care, sundowning. Understanding these symptoms requires more than recognizing the behavior itself; it also requires knowing what happened beforehand. The same behavior may call for a different response depending on its trigger. Video-language models (VLMs) could potentially support caregivers, yet no existing benchmark evaluates this capability. To fill this gap, we present DementiaCare-Bench: 56 professionally produced caregiver training videos segmented into 94 clips across nine BPSD categories, with 2023 questions generated by a multi-agent pipeline that grounds every clinical claim in a verbatim transcript span. Each question is then probed under four visual conditions and labelled by the least it requires, so its visual demand is measured rather than assumed. Measurement contradicts intent: we wrote 77.7% of the questions to require ordered frames, and 34.8% do. Across 12 current VLMs the pattern is uniform. The best reach 85% overall, but that average is carried by questions a language model can answer from clinical knowledge alone; accuracy falls by 17 points on average on questions that require the ordered clip, and a leading open model scores at chance on judging whether a caregiver's response was appropriate. A lightweight LoRA fine-tune, DemCare-VLM, moves video dependence from -3.3 to +4.5 points, so what the benchmark exposes can be repaired and not only measured.

In urban search and rescue, articulated tracked robots (ATRs) must traverse structured but contact-rich environments such as stairwells and cluttered building interiors. Reliable autonomy remains challenging because robot-terrain interaction (RTI) is hybrid and discontinuous, and effective flipper-track coordination is difficult to model analytically. We present ASTRIL-MPC, a language-guided neural kinematics model predictive control (MPC) framework for autonomous traversal. A learned kinematics model predicts short-horizon task-state increments from a height sequence and recent trajectories; NMPC plans with multi-objective costs and strict feasibility constraints; and a large language model (LLM) proposes bounded updates to selected weights and bounds through a safety-checked interface with range clipping, rate limiting, and consistency checks. The compiled predictor enables a full control cycle within 100 ms. Across three traversal tasks and a multi-height generalization setting, ASTRIL-MPC improves an aggregate traversal-quality score by up to 71% over a non-adaptive NMPC and by 67% over a PPO baseline, while eliminating measurable collision impacts during descent. These results indicate that combining terrain-conditioned neural kinematics, optimization-based planning, and language-guided adaptation yields data-efficient and robust autonomy for articulated tracked robots. Real-robot trials over four indoor obstacles further demonstrate transfer to contact-rich physical traversal.

Reinforcement learning is now the standard way to train large language model agents on long-horizon tasks, where dozens of interdependent actions precede a single sparse reward. Critic-free, group-relative methods such as GRPO suit this regime, but they broadcast one trajectory-level scalar to every step and cannot say which decision drove the outcome. GiGPO recovers a step-level signal by grouping time steps that share an anchor state, yet it merges the step- and episode-level estimates under one fixed weight, spending the same resolution on a pivotal branching decision as on a routine, near-deterministic transition. We argue that the right resolution is state-dependent, and propose GACA, a critic-free estimator whose granularity follows an uncertainty-based criticality proxy. GACA scores every step by the negative log-likelihood its own rollout already records, then blends the two advantages with a per-step weight that grows with that score, so the gradient places more weight on the fine-grained signal at above-average NLL and on the episode-level signal below it. We derive an exact risk decomposition for the implemented mixture and show that sufficiently small modulation improves on fixed mixing under positive directional alignment. A separate conditional result bounds local action-value variation using expected NLL, while an error-projection analysis characterizes when mixing adds value beyond scalar uncertainty reweighting. On ALFWorld and WebShop, GACA improves task success over GRPO and GiGPO at both 1.5B and 7B scales.

Comprehensive benchmark suites are essential for improving large language models (LLMs), but many widely used benchmarks are redundant, making evaluation unnecessarily expensive. Although recent benchmark compression methods (BCMs) can mitigate this cost, many strong BCMs rely on large collections of per-sample evaluation results from numerous LLMs to identify representative samples. Building such collections is also expensive unless they are already public, making these methods difficult to extend to newly released benchmarks. To address this challenge, we present ZipBench, a simple and low-cost BCM with theoretical error and rank-consistency guarantees. ZipBench evaluates only a small set of anchor LLMs, synthesizes pseudo evaluation results to broaden coverage, learns compact sample representations, and selects a small yet representative subset. Building on it, we create ZipBench Zoo, a collection of compact versions of 100+ benchmark proxies spanning text, multimodal, and agent tasks. These benchmark achieve mean absolute errors of 0.002--0.02 and average Spearman correlations of ~0.98 with the full benchmarks. Overall, ZipBench reduces the cost of both LLM evaluation and compact benchmark construction, lowering the barrier to broad LLM research for compute-constrained researchers. The code has been released in https://github.com/MilkThink-Lab/ZipBench.

Long video understanding often behaves like a visual needle-in-a-haystack problem: query-relevant evidence is sparsely distributed across long temporal spans, while packing dense frames into a single VLM context incurs context rot and high cost. Existing video agents often rely on query-agnostic offline preprocessing or ad hoc tool sets, which can miss query-specific details and waste computation. In this work, we present VideoXAgent, a purely online video-agent harness for long video understanding that starts from the given video file and user query, plans and decomposes the task, invokes specialized expert tools on demand, and aggregates multimodal evidence to produce a final answer while resolving conflicts among observations. To support this on-demand invocation, we design a suite of heterogeneous expert tools guided by a data-driven taxonomy of atomic capabilities, spanning scripts, VLMs, and domain models (e.g., detection, OCR, ASR, face recognition). The harness further enforces objective evidence prompting and budget-aware control to curb hallucination and non-termination. Across Video-MME-Long, LongVideoBench-Long, LVBench, and MINERVA, VideoXAgent is competitive with frontier LMMs and video agents under a smaller context footprint---about 50k tokens of agent context per sample, even on hour-long videos. In particular, on complex video-reasoning benchmarks such as MINERVA, it matches this level while using only about 15% of the context of a 1,024-frame dense-packing baseline. Notably, the harness remains effective with a visually weak or even text-only orchestrator, suggesting that strong long-video understanding can emerge from progressive agentic evidence seeking rather than from packing the full video into a single context. Project page: https://go-agent-x.github.io/video_agent_harness/

LLM-based agents rely on heterogeneous interaction capabilities to accomplish complex tasks. Existing approaches often distribute these capabilities across multiple LoRA adapters, which increases adapter storage requirements and introduces routing overhead during inference. A single LoRA avoids this overhead, but learning from diverse agent trajectories under a fixed rank budget presents two challenges. First, trajectories with different interaction traces and parameter gradients can induce equivalent changes in decision distributions, causing repeated updates to overemphasize redundant behavioral changes. Second, an aggregated update may exceed the rank budget of the adapter, and approximating it in weight space can distort the decision changes that it is intended to produce. We propose BQ-LoRA, a low-rank adaptation framework that organizes trajectory updates through a local behavior quotient manifold. It contains two modules, i.e., behavior quotient balancing (BQB) and decision preserving compression (DPC). BQB constructs the quotient manifold from decision distributions and reweights trajectory update directions according to their local density in the quotient tangent space. DPC projects the balanced gradient onto the intrinsic fixed rank tangent space and refactorizes the resulting target by jointly controlling effective weight error and distortion of decision distributions. Experiments on AppWorld and BrowseComp-Plus compare BQ-LoRA with standard LoRA and recent low-rank adaptation methods, while separate ablations evaluate the complementary contributions of both components.

Large language models (LLMs) have achieved strong performance on a wide range of natural language tasks, and recent benchmarks suggest that they are increasingly adept at multi-hop reasoning. However, these benchmarks are typically short-horizon, requiring only a small number of retrieval or inference steps, and provide limited evidence of reliability on real-world tasks that involve following manuals spanning hundreds of pages with complex, interdependent guidelines. In this paper, we introduce Tasks over Application Manuals (TAM), a benchmark for evaluating long-horizon procedural reasoning. We construct TAM by curating real-world tasks from two domains: ICD-10-CM clinical coding (mapping medical conditions to diagnostic codes) and U.S. federal sentencing (computing crime sentencing guideline outcomes, specifically offense levels), with human-validated labels. Each task requires following an authoritative manual with tens of thousands of rules and executing a sequence of interdependent steps across different sections to produce an exact answer. We evaluate general-purpose prompting approaches, including retrieval-augmented generation, ReAct-style prompting, and an agent-harness baseline on GPT-5, and find that the best exact-match performance remains extremely low: 1% on ICD-10-CM coding and 15.5% on sentencing tasks. These results show that current benchmarks may overestimate LLM reasoning ability and miss a key challenge: reliably following long, rule-based procedures. The complete TAM data and code are publicly available.

Agent skills are reusable units for language-model agents, but their risks emerge through model decisions, user context, tool calls, and execution feedback rather than through stable signatures or a single sandbox run. Existing static, dynamic, and benchmark-style evaluations rarely preserve public evidence that can be inspected, searched, and reused. We present SkillAtlas, a hosted attack trace library that converts private agent-skill security report bundles into reviewed, redacted, and searchable public cases. The library contains 3,014 cases, 6,589 traces, 151,131 steps, 233 affected skills, and 8 risk categories; 42.5% of successful cases first become successful after a non-success initial round, and trajectory-grounded labels improve pre-execution guard accuracy to 0.770.

In this work, we present ZGCM-1, a fully open 7B dense foundation model trained from scratch with extreme data, system, and algorithmic efficiency. ZGCM-1 is founded on a core premise: compact models cannot passively memorize the open web, but can overcome parametric capacity limits by coupling deliberate internal thinking with active external tool use. To support this paradigm across a 256K context, we develop an end-to-end, high-efficiency open training recipe: Architecture & System Co-design: interleaved gated sliding-window and full attention, and a stable FP8 Muon optimizer; Progressive Curriculum & MDP Mid-Training: context scaling across 16K, 64K, and 256K, and the reformulation of interaction traces into Markov Decision Processes. Furthermore, we establish an AI-native R&D workflow where agent swarms autonomously manage cluster operations, data curation, and rapid diagnostic evaluation. Extensive evaluations show that ZGCM-1-7B is competitive across 7B model family on general benchmarks. On several challenging mathematical reasoning and agentic search suites, it remains competitive with frontier models orders of magnitude larger, such as Qwen3-235B-A22B and GLM-5.1. We also show that our pre-training design offers a ~4.2x efficiency improvement in 16K pre-training time-to-loss. Across the full development lifecycle, we distill eight actionable empirical findings-spanning architectural scaling, SFT quality pruning, long-context generalization, and agentic co-training dynamics. To facilitate community research, we open-source model weights from the pre-training, mid-training, and post-training stages, intermediate checkpoints, training code, per-stage data and data recipes, and W&B logs.

LLM judges are increasingly used to evaluate and improve AI-generated outputs, yet their reliability for complex professional work remains unclear. We study this problem through Vibe Patenting, an end-to-end patent-drafting testbed for AI-agent evaluation. A separately-invoked LLM judge evaluates generated patent drafts and provides structured feedback for iterative revision. Across multiple inventions and drafting-agent configurations, judge-guided revision consistently improves judge-assessed quality, while unguided revision tends to saturate. Notably, iterative judge feedback enables a low-reasoning agent to approach the performance of a substantially more expensive high-reasoning agent. Stronger models and increased reasoning generally improve judge-assessed drafting quality, while domain-specific agentic workflows provide further gains. We validate the judge against independent evaluation by a professional patent attorney and find meaningful but strongly metric-dependent agreement and systematic calibration differences. These results highlight both the utility and limitations of LLM judges as evaluators and optimization signals for complex professional workflows.

The increasing deployment of AI agents in long-horizon tasks yields massive execution logs. Diagnosing failures within these records is crucial for reliability, as it transforms outcome-level signals into actionable interventions. The sheer scale of the data renders human review impractical, driving the need for automated root-cause attribution (RCA). However, automated RCA methods using LLMs suffer from low diagnostic accuracy, especially as execution traces grow larger. They struggle because relevant information is often sparse, distributed across distant actions, and disconnected from the visible failure, reducing root-cause attribution to a massive search problem. Existing RCA methods typically rely on one-shot LLM judgments to diagnose failures from execution traces. While effective for shorter trajectories, these judges tend to settle on a plausible diagnosis early, leaving critical evidence in longer traces unexamined. We introduce Continual Search, an iterative framework that nudges the judge, over successive turns, to keep searching for unresolved diagnostic evidence. We evaluate Continual Search across four existing RCA benchmarks. Recognizing the lack of massive execution traces in current benchmarks, we introduce MegaRCA-Mix to evaluate RCA at scale. MegaRCA-Mix provides a challenging testbed of 50 human-annotated failure trials spanning long-horizon, execution-heavy tasks. Across multiple benchmark suites and model families, Continual Search consistently improves attribution performance. On MegaRCA-Mix, for example, it improves GPT-5.5's F1 score by more than 40%, from $0.349$ to $0.498$. More interestingly, within the same model family, lower-tier models can even surpass their higher-tier counterparts, demonstrating that effective search supersedes raw model scale.

The strong performance of AI Agents across an impressive variety of tasks is driving an unprecedented investment in agentic infrastructures, however the cost of processing tokens is fast increasing. Web agents automate the execution of web-application tasks described in natural language, by analyzing the web-application's user interface (UI) and interacting with it. This work introduces OdoBot, a novel web-agent architecture that completes tasks at a fraction of the cost when compared to conventional web agents. This is achieved by leveraging a behavioral model of the underlying application constructed by analyzing successful task-execution demonstrations. Our experiments with 45 tasks on the Canvas Learning Management System (LMS) demonstrate that OdoBot uses 44% and 80% fewer tokens than two state-of-the-art competitor agents (Agent-E and WebVoyager), while also surpassing WebVoyager in terms of task success rate.

LLM agents are predominantly benchmarked on short, single-task trajectories, yet real deployments run for hours under contention, surfacing a different class of failures. We use the Clinical Environment Simulator (CES), in which an agent manages an entire emergency-department shift under continuous time and resource pressure, as a testbed: long-horizon execution failures manifest measurably in a single rollout under structured, multi-dimensional grading. On CES, current agents reach the correct diagnosis in most cases yet fail to deliver complete and timely critical actions, revealing an execution gap. We attribute this gap to three long-horizon failure modes, each operationalized as a per-trace counter: instruction-adherence drift, treatment incompleteness, and a severity-equity gap in timeliness. We then introduce Asclepius, an adaptive agent scaffolding with a self-evolving harness that rewrites the operating manual between shifts from trace-level feedback, an externalized clinical skills library for high-stakes regimen knowledge, and three isolated subagents that partition per-turn decisions across the patient queue. On held-out batches never observed during harness evolution, Asclepius improves critical-action correctness by 22% (p = 0.024) over a strong baseline agent framework while preserving diagnostic accuracy, with consistent gains across five LLM judges from three model families; on the full ten-batch set, improvements reach 25% on critical actions and 13% on timeliness. The three failure modes form a coupled bottleneck: decisive reductions appear only when all three components act together.

A clinical agent benchmark can report the same verdict on identical inputs while the agent files a materially different order on each run. Such agents order tests, request medications and place referrals, yet benchmarks typically score one run per task and rarely ask whether identical inputs produce identical actions; MedAgentBench, the benchmark we use, scores a single attempt and says so. To measure this gap we introduce "same-input rerun", which replays a task with every input held fixed and compares the orders rather than the score, with six reliability metrics, and apply it to 1000 MedAgentBench runs across 50 tasks from its five write-capable families, two open-weight models below ten billion parameters quantised to four bits, and two temperatures. The study establishes that action-level divergence exists and can pass unrecorded by the score, not that any rate generalises. Under the 8B model at temperature 0.7, all 43 ordering groups emit a different set of orders across five identical runs, 26 emit the order on some runs and not others, and 28 record a different coded value, dose or analyte. In 22 of those 43 the benchmark reports the same failing verdict for materially different behaviour, as it does for all 10 divergent groups of the 4B model at 0.7. Orders also reach different endpoints across runs, one of which the record server rejects while the agent is told it succeeded. These findings motivate repeated-run evaluation, action-level stability reporting and execution-faithful environment feedback in clinical-agent benchmarks.

While multimodal reasoning has advanced rapidly, solving complex geometry problems critically hinges on active visual assistance, such as constructing auxiliary lines, spurring the rise of Visual Chain-of-Thought (VCoT). However, existing evaluations typically assess visual generation quality and final answer accuracy in isolation, failing to examine whether intermediate visual aids are geometrically valid, effectively utilized in subsequent reasoning, or causally responsible for task success. To bridge this gap, we introduce GeoVAD-Bench, a diagnostic benchmark that pairs a fine-grained five-dimensional trajectory diagnosis covering perception, auxiliary quality, utilization, deductive reasoning, and final correctness with controlled No-Aux, Auto-Aux, and GT-Aux intervention settings to systematically isolate intermediate error modes, the causal gains of visual aids, and the resulting autonomy gap. Our findings reveal that while high-quality auxiliary aids offer substantial theoretical gains for geometric problem solving, autonomous generation is frequently hampered by compounding errors across geometric perception, faithful visual manipulation, visual-state grounding, and deductive reasoning. Guided by these diagnostic insights, we establish a specialized data construction pipeline encompassing geometric perception, diagram editing, and interleaved visual-textual reasoning trajectories, and develop a progressive SFT and multimodal RL training framework. The resulting model, GeoWeave-8B, outperforms the base model by +25.3% in final geometric accuracy and achieves a +30.4% gain in process average across the four intermediate diagnostic dimensions.

Large language model agents are expected to continuously adapt to new tasks and environments over their lifetime by reusing past experience. However, existing memory-based agents struggle to transfer reusable experience across environments and suffer from catastrophic forgetting as experience accumulated. To address these challenges, we propose LifeMem, a lifelong learning framework that enables agents to transfer knowledge across multiple environments. During learning, LifeMem clusters accumulated interaction trajectories based on underlying workflows to extract reusable skills. When solving a new task at inference time, the agent recalls relevant skills and trajectories to guide actions. To validate our method, we conduct experiments across 10 environments and over 13k tasks with 2k newly annotated interaction trajectories. Results show that LifeMem enables effective experience reuse in lifelong learning, achieving both reduced forgetting on learned tasks and superior cross-task transfer. Further analysis reveals that task streaming impacts learning, while consolidating structurally similar trajectories within memory boosts performance.

AI agents can automate tasks by turning a single natural-language request into a multi-step process spanning tools, files, and applications. Users are often left to judge that process from fragmented execution information and the final output. To make the completed process easier to understand, validate, and reuse, we investigate post-task workflows: editable, graph-based representations of an agent's completed execution. We first analyzed 10,803 public workflow templates from n8n to characterize real-world automation practice, then developed Trace2Flow, a research probe that translates agent execution traces into interactive post-task workflows. In a study, participants (N = 20) reviewed agent executions with prompt or agent errors. We found that post-task workflows improved their understanding and error detection over a prompt-only condition, and that validation succeeded mainly when users cross-checked across multiple evidence sources. For follow-up tasks, adapting the workflow matched adapting the prior prompt in success, time, and difficulty, and was often preferred.

We demonstrate that a general-purpose agent can directly drive a physical robot throughout task execution without any task-specific or environment-specific training. We introduce Agent as Policy (AGP), which places task planning and execution under the agent's control. Given a task and a robot interface, the agent interprets visual evidence, writes executable programs, issues motion commands, and revises its actions in response to physical outcomes. This brings the agent's reasoning and programming capabilities into continuous interaction with the physical world. We study AGP across multiple real-world manipulation tasks spanning precision manipulation, dynamic motions, and deformable objects. These include assembly from human videos, block construction from goal images, dice flipping, targeted throwing, and bimanual towel folding. Across assembly, block construction, and dice flipping, AGP succeeds in at least eight of ten trials for each evaluated task configuration. We further study efficiency through task experience accumulation and find that reusing saved procedures and programs shortens execution time across repeated trials. These findings support a path for general-purpose agents to act as robot policies, extending their autonomy to physical manipulation through runtime reasoning, programming, and interaction.

Mobile GUI agents are shifting from multi-module frameworks to native models trained end-to-end, yet industrial deployment faces three persistent gaps. Sandbox training produces a distribution mismatch with production environments; expensive real-device failures remain underutilized; and fixed benchmarks saturate, losing the power to guide iteration. We present BlueLM-GUI, a 35B-A3B mobile GUI agent built as a real-device-centric flywheel that closes these gaps through three principles. Every Sample Matters: a dual-track pipeline with Heterogeneous Triple-System Consensus evaluation and an Error Correction & Derivation Module salvages every trajectory into usable supervision. Every Rollout Is Real: a three-stage recipe---continual pre-training, supervised fine-tuning, and agentic reinforcement learning on hundreds of real phones---grounds every rollout in real production environments, so the capability the model learns transfers directly to deployment. Every Query Evolves: a quota-driven benchmark methodology with three orthogonal axes enables precise attribution and allows the benchmark to be systematically upgraded as the model improves. BlueLM-GUI achieves 87.4 on MobileGUI-VBench, surpassing the best closed-source model by 5.1 points, and 84.9 on AndroidWorld, the best result among open-source models and competitive with closed-source models. These results demonstrate that grounding model training and iterative improvement in both real devices and the three Every principles yields strong, robust, and transferable mobile GUI capability.

Web agents can utilize reusable tools to reduce the cost and latency of low-level browser interaction, but automatically discovered tool collections can be large, redundant, and poorly aligned with user demand. We present AutoTailor, a meta-agentic framework for constructing and maintaining a compact set of trajectory-derived Model Context Protocol (MCP) APIs. Offline, AutoTailor converts web trajectories into parameterized browser-automation programs, applies a Quality Filter to remove APIs with unsuitable granularity and redundant functionality, and applies a Usage Likelihood Filter to prioritize broadly useful capabilities while preserving semantic coverage. Online, Dynamic Reselection monitors task outcomes and API usage, identifies recurring coverage gaps, adds relevant candidates, and prunes persistently unused capabilities. We evaluate AutoTailor on 106 WebArena Postmill tasks. Offline filtering reduces the initial 1,283 unrefined APIs to 87, and Dynamic Reselection produces a 33-API set. With reasoning and acting (ReAct) fallback, this set achieves 90.6% correctness, compared with 87.5% for ReAct alone, while reducing average total request-token cost by 57.8% and latency by 29.4%. Without ReAct, it achieves 60.1% correctness, marginally matching the performance of unrefined set, while reducing request-token usage by 94.9%. Together, these results show that static filtering produces a compact inventory of APIs expected to support core, high-likelihood tasks, while dynamic reselection further tailors that inventory to observed user needs. This combination improves accuracy and latency while sharply reducing token usage and end-to-end cost, demonstrating the value of user-aligned capability management for efficient web agents.

Enterprise AI agents often succeed in a demonstration and then stall once they must operate day after day. An industry report estimates that most pilots never reach production and that deployed systems rarely retain feedback or improve over time, while agent benchmarks show single-run successes masking unreliable repetition. We argue that these failure modes share a common architectural root: agents reason over data structured for human operators and traditional applications, not for the language models that power them. This position paper proposes that companies deploying agents in sustained operation should rebuild their cognitive substrate, the shared environment agents read as working context, around representations matched to that reasoning surface, isolating schema translation to the action boundary. Markdown is the instantiation available today, not a proven agent-native primitive. Two mechanisms ground the argument: context-bandwidth asymmetry, the gap between one-pass reading of connected prose and field-by-field typed access that strips relations; and cross-loop coupling, the claim that action, skill, and policy loops compound only if they share one substrate. A four-layer framework (Data, Knowledge, Intelligence, Governance) operationalizes the position, with a Sync Agent enforcing the action boundary and a per-skill trust gradient, making governance and auditability structural properties of the substrate. The position revives the shared-substrate tradition of classical multi-agent systems under LLM-era economics. We analyze the main objections and risks, including indirect prompt injection on the compile path, and outline a research agenda for evaluating substrates directly.

Large language models (LLMs) are rapidly evolving from conversational assistants into agentic AI systems that reason, plan, invoke tools, maintain persistent memory, communicate with other agents, and execute multi-step tasks. At the same time, modern models exhibit substantially stronger native safety alignment than earlier generations on which many jailbreak attacks and defenses were originally studied. This shift raises a fundamental question: which established jailbreak-security findings remain valid in the era of modern LLMs and agentic AI? We address this question through a Systematization of Knowledge (SoK) that reframes jailbreak security around the full agentic execution pipeline. We develop unified taxonomies of attacks and defenses spanning user interaction, planning and reasoning, memory, tool use, and inter-agent communication, and introduce a security--utility--efficiency evaluation framework that separates native harmful-prompt safety, adversarial jailbreak robustness, and agent-level security outcomes. We further conduct a controlled empirical study of representative attacks and defenses within a common agentic framework. Our results reveal three important gaps. First, strong native alignment does not imply robustness to adversarial jailbreaks. Second, defense effectiveness is highly model-, attack-, and component-dependent and can come at substantial cost in over-refusal, utility, and latency. Third, low final-response attack success can mask severe intermediate compromise: planning, memory, and tool interactions may remain unsafe even when the final response is successfully filtered. These findings motivate a shift from response-centric jailbreak defense toward cross-layer, execution-aware security that protects agent state, component transitions, and external actions while preserving practical utility and efficiency.

Large language models are being organized into multi-agent systems with specialized roles, but whether such specialization produces distinct forecasts and whether subsequent synthesis improves utility remains unclear. In this study, we carried out a live, prospective evaluation over the final 56 matches of the information-dense 2026 FIFA World Cup, keeping a frontier foundation model constant while assigning two primary forecasting agents contrasting specialist roles: a quantitative specialist focusing on structured performance statistics and a news specialist focusing on current injuries, tactics and information from press conferences. Their forecasts were then reviewed by a separate critic before being combined by a meta-agent, resulting in a sequential four-agent model. Forecasts from the betting market served as an external benchmark. The news specialist obtained the highest mean probability-weighted Top-3 utility and matched the betting market in Top-3 exact-score hits. Nevertheless, the two specialist forecasters agreed on at least two of the three scorelines in 50 out of 56 matches, and the meta-agent never generated more than one scoreline outside the specialists' forecast set. These findings show that rapidly changing, unstructured information can provide a valuable forecasting signal alongside structured statistics, whereas adding critic and meta-agent stages does not necessarily create complementary information or improve on the strongest specialist.

Existing agent benchmarks mainly evaluate final task success or tool-call correctness, providing limited insight into whether agents can reliably diagnose and recover from intermediate execution failures. This limitation becomes particularly critical in multi-turn parallel tool-use scenarios, where errors may propagate across dependent branches and trigger cascading failures. We introduce ParaRecover, a process-level benchmark for evaluating error localization and recovery in multi-turn parallel tool-use agents. Built upon a fine-grained taxonomy of 14 error types covering planning dependencies, tool selection, and argument matching, the benchmark comprises 10,626 instances spanning two difficulty levels. To enable finegrained, process-oriented evaluation, we further propose the SDE rubric, which measures structural integrity, diagnostic reasoning, and evolutionary strategy during agent execution.Experiments across more than ten mainstream LLMs reveal that even state-of-the-art models still struggle with multi-turn error propagation,implicit tool-use failures, and precise replanning. Moreover, we demonstrate that the SDE rubric provides effective supervision signals for improving agents' reflective recovery capabilities. Our data and code are available at https://github.com/gbw206/ParaRecover.

Lux AI Season 3 requires agents to act under partial observability, randomized episode level dynamics, and a best of five match structure that rewards both tactical execution and fast adaptation. We present HORIZON, a hierarchical agent that combines symmetry aware spatial perception, dual memory belief tracking, relic centric graph attention, information gain driven exploration, and an opponent conditioned policy mixture. HORIZON separates short horizon control from cross match meta reasoning, while auxiliary belief and world model objectives stabilize learning. Trained with PPO in a large scale JAX simulator, the resulting agent explicitly infers hidden game parameters and opponent style. Experiments show consistent gains in match win rate, episode win rate, adaptation gain, and league rating over strong recurrent and feed forward baselines.

Large Language Models (LLMs) have enabled more natural human-robot interaction, but open-source models often exhibit unstable long-horizon reasoning and inefficient action execution when deployed in agentic robotic frameworks. This paper presents an enhanced ROS-Agent based architecture that improves task reliability and execution efficiency for agentic robotic systems using open-source LLMs. The proposed system introduces a novel intermediate mechanism, termed the MetaTool, which enforces structured planning prior to action execution. Given a natural-language command, the MetaTool induces the LLM to generate a pseudo-code plan of intended tool invocations, which is stored in the ROS-Agent's scratchpad and persists throughout execution. By explicitly separating planning from execution, the proposed approach reduces execution loops and improves deterministic behavior. The architecture is validated on a custom mobile robotic platform with multimodal perception and motion control capabilities. Experimental results on real-world interactive tasks demonstrate improved task completion and contextual consistency, with up to ~24% gains on complex tasks compared to the baseline framework.