Multimodal Agents

Ask about this topic

Overview

Multimodal agents combine language models with visual, audio, and video perception to answer questions, use tools, and interact with apps, websites, robots, or other environments. Common approaches include planning and tool orchestration, specialized collaborating agents, active evidence gathering, and memory or reusable skills for long tasks; reinforcement learning and execution feedback are increasingly used to improve these behaviors. Benchmarks now test browsing, GUI use, video reasoning, and embodied tasks, often evaluating action traces and process reliability as well as final answers. Despite progress, agents remain brittle over long horizons, and more modalities or more agents do not automatically improve grounding, safety, efficiency, or task success.

Latest papers 367

49 papers in the last four weeks, up 20% on the four weeks before. 0.3% of all new papers.

All topics
CardsList
Jun 10, 2026

MedCTA: A Benchmark for Clinical Tool Agents

To make clinically grounded decisions, medical AI agents are expected to go beyond simple recognition and be capable of tool retrieval, evidence acquisition, and integration. Existing benchmarks largely evaluate isolated perception or single-turn question answering, and therefore provide limited visibility into failures of planning, tool recruitment, and rollout reliability. We introduce MedCTA, a benchmark for evaluating medical tool agents on clinician-validated, step-implicit tasks grounded in realistic multimodal clinical inputs, including radiology images, pathology slides, and reports. MedCTA comprises 107 real-world clinical tasks with clinician-verified executable trajectories over 5 deployed tools, and supports process-aware evaluation of tool selection, argument validity, execution stability, trajectory fidelity, and outcome quality. We benchmark 18 open- and closed-source multimodal models and find that even frontier systems remain brittle in multi-step clinical tool use: autonomous rollouts are dominated by protocol failures, premature stopping, and incorrect tool recruitment, while gold-standard tool routing yields large but still incomplete gains. These results show that strong backbone perception does not translate into reliable agentic behavior in clinical settings. MedCTA provides a rigorous testbed for auditing, diagnosing, and advancing trustworthy medical AI agents. The dataset and evaluation suite are available at https://ivul-kaust.github.io/MedCTA/

Tajamul Ashraf, Hyewon Jeong, Fida Mohammad Thoker +1

King Abdullah University of Science and Technology (KAUST), Saudi Arabia · Massachusetts Institute of Technology (MIT), USA

Agent EvaluationAI Agent BenchmarksTool-Using AgentsMultimodal Agents
AskPDF
Jun 10, 2026

Orchestra-o1: Omnimodal Agent Orchestration

The recent success of agent swarms has shifted the paradigm of large language model (LLM)-based agents from single-agent workflows to multi-agent systems, highlighting the importance of agent orchestration for task decomposition and collaboration. However, existing orchestration frameworks are limited to a narrow set of modalities and struggle to generalize to more complex settings where heterogeneous modalities coexist and interact. This limitation becomes particularly pronounced in omnimodal scenarios, where tasks require the unified understanding and coordination of diverse inputs such as text, image, audio, and video. In this work, we propose Orchestra-o1, an omnimodal agent orchestration framework designed to support efficient agent collaboration across multiple modalities. Orchestra-o1 introduces a unified orchestration mechanism that enables modality-aware task decomposition, online sub-agent specialization, and parallel sub-task execution. This scalable design allows agent systems to effectively tackle complex real-world tasks involving heterogeneous information sources, surpassing the second-best approach by 10.3% accuracy on the OmniGAIA benchmark. Furthermore, we introduce decision-aligned group relative policy optimization (DA-GRPO), an efficient agentic reinforcement learning approach for training Orchestra-o1-8B, which also achieves state-of-the-art performance against all existing open-source omnimodal agents.

Fan Zhang, Vireo Zhang, Shengju Qian +8

CUHK · LIGHTSPEED · PKU +2

LLM Agent OrchestrationGroup Relative Policy OptimizationMultimodal AgentsRL for LLM Agents
AskPDF
Jun 10, 2026

IAPO: Input Attribution-Aware Policy Optimization for Tool Use in Small Multimodal Agents

This paper investigates reinforcement learning (RL) methods for improving tool-calling capabilities in multimodal small language model (SLM) agents. While existing works have explored various reward designs to improve agentic tool-calling ability, these approaches face inherent limitations for SLM training, especially under multimodal scenarios. First, many existing methods evaluate tool use correctness through exact matching against certain ground-truth or predefined formats. However, this assumption is often unsuitable for multimodal tasks, where multiple tool use paths may be valid and annotated tool trajectories are typically unavailable. Second, such sparse and brittle binary rewards provide little guidance on how to improve the underlying decision process, making them particularly difficult for multimodal SLM to learn from. To address these issues, we propose Input Attribution-Aware Policy Optimization (IAPO), an RL algorithm for improving tool use in multimodal SLM by aligning the model's attribution across input components with that of a stronger teacher. Experiments on Qwen2.5-VL-3B show that the proposed method improves visual question answering accuracy by an average of 3% across six test sets compared with existing visual tool use work, by helping the model attend to the most relevant input evidence.

Yifan Yang, Zhen Zhang, Jiayi Tian +2

University of California, Santa Barbara

Tool Use in VLMsReinforcement LearningMultimodal QAMultimodal AgentsRL for Tool Use
AskPDF
Jun 9, 2026

Data Journalist Agent: Transforming Data into Verifiable Multimodal Stories

Data tells stories that shape society; the data journalist's job is to turn raw information into stories non-experts can trust. A high-quality news feature takes a newsroom team weeks: hunting for context, running statistics, choosing an angle, and designing visuals. Recent agents handle individual steps well: data-science agents close the analysis loop, while design agents synthesize beautiful websites. But can an agent serve as a data journalist end to end? We introduce Data Journalist Agent (Data2Story), a multi-agent framework that orchestrates specialized roles into a single virtual newsroom. Data2Story contributes two innovations. (i) Claims are evidence-grounded: an Inspector links every number, angle, and asset back to data, code, or an external reference. (ii) Articles are multimodally generative: rather than defaulting to plain text and static charts, Data2Story reasons about what readers will want to see, then deploys multimodal tools, such as interactive maps for geography and audio for music. We evaluate Data2Story on 18 articles, each paired with the originally published expert piece, along four axes: (a) human-agent angle coverage; (b) rubric evaluation with 53 participants across five dimensions; (c) computer-use agents as judges, a cost-saving proxy for how readers navigate interactive articles; and (d) verifiability, where a coding verifier re-executes statements against the data and checks claims against references. Data2Story produces competitive, evidence-traceable multimedia stories, with particular strength in transparency and auditability. Human articles retain an edge in editorial angle, creative design, and presentation. We position Data2Story as a collaborator for journalists, enabling more evidence-based, transparent, and verifiable reporting. Code and demos are available at https://data2story.github.io.

Kevin Qinghong Lin, Batu EI, Yuhong Shi +5

University of Oxford · Stanford University · BBC R&D

LLM Agent OrchestrationMultimodal AgentsMultimodal Tool Use
AskPDF
Jun 9, 2026

MemVenom: Triggered Poisoning of Multimodal Memories in Web Agents

External memory has become a core component of modern web agents, enabling long-horizon reasoning through the retrieval of past experiences. However, this paradigm introduces a critical vulnerability: malicious content injected into memory can be persistently recalled and repeatedly influence agent behavior. In this work, we identify and systematically study multimodal memory poisoning, an overlooked yet practical attack surface in web-agent systems. We propose MemVenom, a unified black-box attack framework that poisons graph-structured external memory with coordinated text-image evidence. Our method consists of a two-stage design: (1) a trigger-conditioned retrieval attack that ensures high-probability recall of malicious memory, and (2) a post-retrieval attack induction that leverages adversarial perturbations and stealthy OCR injection to override the original user objective. Unlike prior attacks that operate on prompts or text-only memory, our approach enables persistent, reusable, and goal-agnostic attacks without modifying model parameters or re-optimizing malicious tasks. Experiments across multiple web-agent frameworks and vision-language models demonstrate that MemVenom achieves strong end-to-end attack success with minimal impact on benign performance, reaching up to 99.15% on GPT-5-family web agents, while transferring effectively across architectures and model scales.

Yv Zhang, Hao Sun, Hao Fang +5

1Harbin Institute of Technology, Shenzhen · 2Peng Cheng Laboratory · 3Tianjin University +2

Agent Memory PoisoningBlack-Box Adversarial AttacksMultimodal AgentsWeb Agents
AskPDF
Jun 9, 2026

Kwai Keye-VL-2.0 Technical Report

We introduce Kwai Keye-VL-2.0-30B-A3B, an open-source Mixture-of-Experts (MoE) multimodal foundation model designed to advance long-video understanding and agentic intelligence. To address the challenges of ultra-long contexts, information redundancy, and prohibitive computational costs inherent in hour-level videos, Keye-VL-2.0 is the first to adapt DeepSeek Sparse Attention (DSA) to GQA-based multimodal architectures, enabling lossless 256K context processing while capturing critical frames and long-range temporal dependencies. This architecture is underpinned by a highly optimized training and inference infrastructure, including scalable video I/O, heterogeneous ViT-LM parallelism, and custom DSA kernels that significantly maximize throughput and minimize computational overhead. Furthermore, to overcome the algorithmic dilemma of catastrophic forgetting during multi-task alignment, we introduce Cross-Modal Multi-Teacher On-Policy Distillation (MOPD) paired with Context-RL and Video-RL. By distilling dense token-level teacher feedback from on-policy rollouts back into the MoE backbone, which activates only 3B parameters, Keye-VL-2.0 natively empowers advanced agent collaboration across Code, Tool, and Search scenarios with multimodal self-correction. Extensive evaluations across video understanding, temporal grounding, reasoning, STEM, and agent benchmarks demonstrate that Keye-VL-2.0-30B-A3B achieves state-of-the-art performance among models of similar scale, particularly excelling in fine-grained temporal localization on TimeLens and long-video comprehension on Video-MME-v2 and LongVideoBench. We release our model checkpoints to accelerate community progress toward scalable and robust multimodal agentic applications.

Kwai Keye Team, Bin Wen, Changyi Liu +50

Temporal Video GroundingLong-Context ModelingOn-Policy DistillationMultimodal Foundation ModelsMultimodal AgentsTemporal Video Understanding+1
AskPDF
Jun 8, 2026

ReasonRec: A Reasoning-Augmented Multimodal Agent for Unified Recommendation

Recent advances in multimodal recommenders excel at feature fusion but remain opaque and inefficient decision-makers, lacking explicit reasoning and self-awareness of uncertainty. We introduce ReasonRec, a reasoning-augmented multimodal agent structured around a three-stage explicit reasoning pipeline. Specifically, we propose a reasoning-aware visual instruction tuning strategy that systematically transforms diverse recommendation tasks into unified CoT prompts, enabling the VLM to explicitly articulate intermediate decision steps. Additionally, our evidence-horizon curriculum progressively enhances the reasoning complexity to better handle cold-start and long-tail user scenarios, significantly boosting model generalization. Furthermore, the uncertainty-guided delegation mechanism empowers the agent to assess its own confidence, strategically allocating computational resources to optimize both recommendation accuracy and inference efficiency. Comprehensive experiments on four standard recommendation tasks across five real-world datasets demonstrate that ReasonRec achieves over 30% relative improvement in key ranking metrics compared to state-of-the-art multimodal recommenders. Crucially, ReasonRec substantially reduces inference latency by dynamically delegating up to 35% of queries to efficient sub-models without compromising accuracy. Extensive ablation studies further confirm that each proposed reasoning and planning mechanism individually contributes substantially to ReasonRec's overall effectiveness. Collectively, our results illustrate a clear pathway towards interpretable, adaptive, and efficient multimodal recommendation through explicit reasoning and agentic design.

Yihua Zhang, Mingfu Liang, Jiyan Yang +11

Meta AI · Michigan State University · The University of North Carolina at Chapel Hill +1

Efficient Multimodal InferenceMultimodal CoT ReasoningRecommender SystemsLarge Language Model-Based RecommendationMultimodal RecommendationMultimodal Agents+1
AskPDF
Jun 8, 2026

SpatialWorld: Benchmarking Interactive Spatial Reasoning of Multimodal Agents in Real-World Tasks

Spatial reasoning is a foundational capability for multimodal large language models (MLLMs) to perceive and operate within the physical world. However, existing benchmarks predominantly rely on passive evaluation (e.g., static VQA) or simulator-specific pipelines, failing to assess general interactive spatial understanding. We introduce SpatialWorld, a unified benchmark designed specifically for evaluating the interactive spatial understanding of multimodal agents in complex real-world tasks. Integrating eight heterogeneous simulation backends under a shared, simulator-agnostic protocol, SpatialWorld features 760 human-annotated tasks across diverse domains (e.g., household routines, travel, social collaboration). Agents must solve tasks under vision-only partial observability, actively gathering egocentric visual evidence and expressing decisions via a unified, text-based action interface native to MLLMs. For reliable evaluation, each task includes a human-validated initial state, a reference trajectory, and a terminal-state verifier. Evaluating 15 advanced agents reveals that robust spatial task solving remains challenging: the strongest model, GPT-5, achieves an average task success rate (TSR) of only 17.4%, while the leading open-source model, Qwen-3.5, reaches 14.1%. Further analysis exposes a clear mismatch between task success and execution efficiency, alongside substantial domain-specific performance variations. These bottlenecks in active exploration and long-horizon planning position SpatialWorld as a rigorous testbed for future spatial agents.

Hongcheng Gao, Hailong Qu, Jingyi Tang +18

Tsinghua University · Chongqing University · Peking University +6

Spatial Reasoning BenchmarksVisual Spatial ReasoningLong-Horizon Agent EvaluationAI Agent BenchmarksMultimodal AgentsSpatial Reasoning
AskPDF
Jun 8, 2026

HDRAgent: An Agentic Framework for Multi-Exposure HDR Imaging

Most existing multi-exposure HDR methods follow a fixed feed-forward reconstruction paradigm, making them prone to ghosting artifacts in complex dynamic scenes. To address this issue, we propose HDRAgent, the first agent-driven framework for HDR imaging, which adaptively selects reconstruction strategies according to the current scene conditions. Specifically, to provide scene-specific prior knowledge, we introduce a fine-grained contextual knowledge matching (FCM) module. This module leverages multimodal large language model (MLLM)-derived scene perception to retrieve relevant historical cases and tool knowledge, organizing them into structured evidence for MLLM-based adaptive tool scheduling. In addition, we propose a perception--distortion feedback mechanism that transforms post-execution quality assessment and artifact diagnosis into structured feedback, which is accumulated in historical memory to help subsequent contextual knowledge refinement and strategy selection. Furthermore, considering that extreme motion can invalidate alignment methods, we design an agent-guided generative alignment strategy that uses MLLM-based dynamic-region parsing to reconstruct unreliable contents in non-reference frames under reference-frame guidance. Experiments demonstrate that HDRAgent effectively reduces ghosting and local artifacts while achieving competitive or superior objective performance and visual quality.

Weiyu Zhou, Tao Hu, Yijian Wang +3

School of Computer Science, Northwestern Polytechnical University, Xi’an 710129, China · Zhejiang University, Hangzhou, China · Camera Group, DJI, Shenzhen, China +1

Image EnhancementHDR ImagingMultimodal Agents
AskPDF
Jun 6, 2026

A Multi-modal Agentic Co-pilot for Evidence Grounded Computational Pathology

Pathology is the cornerstone of modern medicine, where accurate decision-making relies heavily on evidence-based practices. While artificial intelligence (AI) has the potential to transform clinical workflows, the intersection of AI and evidence-based medicine remains under-explored, with primitive attempts restricted to text-only general medicine. In this work, we present PathPocket, a multimodal AI agentic co-pilot designed specifically for evidence grounded pathology. We construct the most comprehensive pathology evidence corpus to date, encompassing approximately 110,472 public and authorized documents structured across a rigorous hierarchy of evidence from clinical guideline to expert opinion. From this meticulously graded foundation, we build a large-scale multimodal pathology hypergraph containing over 4.55 million entities and 7.10 million relations. Serving as a robust knowledge engine, this hypergraph provides traceable evidence for a collaborative multi-agent reasoning framework integrating input understanding, evidence retrieval, filtering, and diagnosis generation. This enables PathPocket to seamlessly resolve a wide spectrum of clinical tasks, ranging from text-only queries to complex multimodal diagnostics involving region-of-interest (ROI) and gigapixel whole-slide images (WSIs). We rigorously evaluate the system on a multidimensional benchmark of over 200,000 real-world cases, where it significantly outperforms existing state-of-the-arts. Crucially, extensive user studies demonstrate that PathPocket substantially improves the diagnostic accuracy and confidence of pathologists. By directly grounding pathology interpretations in verifiable literature, PathPocket offers a practical and scalable solution for the future of evidence grounded computational pathology.

Zhe Xu, Zhengyu Zhang, Zhiyuan Cai +12

Department of Computer Science and Engineering, Hong Kong University of Science and Technology, Hong Kong, China · Department of Pathology, Nanfang Hospital, Southern Medical University, Guangzhou, China · Department of Pathology, School of Basic Medical Sciences, Southern Medical University, Guangzhou, China +7

Computational PathologyClinical Decision SupportMedical QAMultimodal AgentsKnowledge GraphsWhole-Slide Image Analysis+1
AskPDF
Jun 6, 2026

Aligned but Not Partner-Specific: How Multimodal LLM Agents Succeed in Reference Games Without Forming Conceptual Pacts

Repeated reference games test whether interlocutors replace their initially long descriptions with shorter, partner-specific expressions grounded in shared interaction history; that is, with conceptual pacts. Prior work shows that multimodal LLMs fail to become more efficient across rounds, although they align on the labels they use. However, how can we determine whether this alignment reflects partner-specific grounding rather than a shared task vocabulary? We address this by comparing competent multimodal agent dyads with human dyads from the KTH Tangrams corpus. Our novel methodological contribution is a pragmatically constrained pseudo-dyad baseline: rounds from two different real dyads describing the same target at comparable trajectory positions are paired, preserving referential task structure while removing shared partner history. This enables us to test whether the observed label alignment depends on interaction with a specific partner. Across three measures (task competence, description strategy, alignment dynamics), we find clear differences. Humans reduce effort through entrainment, compressing descriptions and increasing label alignment with partners. Agents instead maintain fixed effort levels, producing verbose descriptions from round one, with near-ceiling label overlap that is statistically indistinguishable between real and pseudo dyads. MLLMs thus achieve coordination without conceptual pacts, succeeding by verbose description rather than by forming the compact, history-dependent referring expressions characteristic of human dialogue.

Po-Ya Angela Wang, Chinmaya Mishra, Aslı Özyürek +2

Co-Intelligence Humanities AI Future Lab and Graduate Institute of Linguistics, National Taiwan University · Multimodal Language Department, Max Planck Institute for Psycholinguistics · Donders Institute for Brain, Cognition and Behaviour, Radboud University +1

Repeated GamesLLM Agent EvaluationMultimodal AgentsMulti-Agent Communication
AskPDF
Jun 5, 2026

Struct-Searcher: Agentic Structural Thinking Advances Multimodal Deep Information Seeking

Deep research agents have attracted increasing attention for their ability to collect large-scale online information to acquire target knowledge, with recent efforts shifting from purely text-based information seeking to multimodal settings. However, existing agentic workflows are largely aligned with evidence accumulation models, which linearly aggregate evidence and lack principled mechanisms for handling contradictory information across heterogeneous modalities. Towards this end, we propose Struct-Searcher, a structural agentic workflow grounded in belief revision theory that explicitly maintains an evolving multimodal structural graph throughout the reasoning process, enabling effective conflict-aware multimodal deep information seeking. Extensive experiments across multiple benchmark datasets and backbone models demonstrate that Struct-Searcher is (1) plug-and-play and model-agnostic, yielding an average relative accuracy improvement of 17.2% on BrowseComp-VL across five different backbones. (2) top-performing, consistently outperforming state-of-the-art vision-language models (VLMs) and deep research agents, with relative accuracy improvements of 3.7% on MM-BrowseComp, 1.5% on HLE-VL, and 0.7% on BrowseComp-VL over the second-best competing approach.

Fan Zhang, Vireo Zhang, Shengju Qian +7

1CUHK · 2LIGHTSPEED · 3PKU +3

Multimodal IRBelief RevisionDeep Research AgentsMultimodal ReasoningMultimodal Search AgentsMultimodal Agents
AskPDF
Jun 4, 2026

Resonant Minds: Closed-Loop Social Avatars with Theory of Mind

Creating lifelike digital humans with genuine social intelligence requires unifying cognitive reasoning and multimodal generation within a coherent framework. Current approaches treat these as separate tasks: Large Language Models excel at dialogue but lack embodied expression, while diffusion-based talking head models achieve visual fidelity but ignore social cognition. To bridge this gap, we propose a closed-loop dual-agent framework integrating perception, social reasoning, and expression into a continuous interaction cycle. The perception module analyzes partners' multimodal behaviors from video, while the social reasoning module infers hidden mental states through Theory of Mind and selects responses via an ensemble mechanism. The expression module then generates emotion-controllable videos that jointly synthesize speaker speech and facial expressions with listener reactive behaviors, capturing bidirectional dynamics absent in prior work. We further construct a hierarchical Persona-Scenario dataset with psychologically grounded personas and private social goals to support evaluation under information asymmetry. Experiments on this dataset demonstrate competitive or superior performance on both dialogue quality and video generation metrics. Notably, our method surpasses even the full-information Script mode on key dialogue quality dimensions, suggesting that explicit mental state inference under uncertainty can elicit more thoughtful dialogue than unrestricted information access. Project page: https://resonantminds.github.io/.

Jianxu Shangguan, Jing Xu, Hang Ye +4

University of Washington · 2Peking University · 3Carnegie Mellon University +1

Audio-Video GenerationTheory of MindMultimodal AgentsEmotion-Aware Dialogue Generation
AskPDF
Jun 3, 2026

Plan, Watch, Recover: A Benchmark and Architectures for Proactive Procedural Assistance

Proactive multimodal assistants must decide when guidance is needed and what to say, while tracking progress through extended procedures and helping users recover when they depart from the expected plan. Progress is limited by benchmarks that score isolated clips, cover narrow domains, or omit realistic deviations. We introduce EgoProactive, a smart-glasses dataset of procedures across four domains in which users commit step-level deviations and recover from them, paired with reference guidance. We combine it with five re-annotated egocentric and instructional datasets to form Pro2^{2}Bench, a cross-domain benchmark that scores intervention timing, procedure completion, and response quality. Zero-shot, no frontier or open-weight model completes more than 26% of a procedure's interventions before its first error. We propose PWR (Plan, Watch, Recover), which keeps an explicit plan state in the model's context and revises it after every intervention, including recovery steps after a deviation. Across three open-weight backbones, training on Pro2^{2}Bench without a plan barely changes performance, while adding the dynamic plan raises average procedure completion from 20% to 27% and improves turn-level decisions and response quality for every backbone. The gains hold on held-out datasets and on deviations never seen in training.

Kaustav Kundu, Ritvik Shrivastava, Maxim Arap +14

Meta Reality Labs · Meta Superintelligence Labs

LLM PlanningEgocentric Video UnderstandingAI Agent BenchmarksProactive AssistanceMultimodal AgentsLLM Agent Planning
AskPDF
Jun 2, 2026

CP-Agent: Context-Aware Multimodal Reasoning for Cellular Morphological Profiling under Chemical Perturbations

Cell Painting combines multiplexed fluorescent staining, high-content imaging, and quantitative analysis to generate high-dimensional phenotypic readouts to support diverse downstream tasks such as mechanism-of-action (MoA) inference, toxicity prediction, and construction of drug-disease atlases. However, existing workflows are slow, costly and difficult to interpret. Approaches for drug screening modeling predominantly focus on molecular representation learning, while neglecting actual experimental context (e.g., cell line, dosing schedule, etc.), limiting generalization and MoA resolution. We introduce CP-Agent, an agentic multimodal large language model (MLLM) capable of generating mechanism-relevant, human-interpretable rationales for cell morphological changes under drug perturbations. At its core, CP-Agent leverages a context-aware alignment module, CP-CLIP, that jointly embeds high-content images and experimental metadata to enable robust treatment and MoA discrimination (achieving a maximum F1-score of 0.896). By integrating CP-CLIP outputs with agentic tool usage and reasoning, CP-Agent compiles rationales into a structured report to guide experimental design and hypothesis refinement. These capabilities highlight CP-Agent's potential to accelerate drug discovery by enabling more interpretable, scalable, and context-aware phenotypic screening -- streamlining iterative cycles of hypothesis generation in drug discovery.

Yuxin Zhang, Yiyao Li, Ping Shu Ho +3

Department of Electrical and Computer Engineering, The University of Hong Kong · School of Computing and Data Science, The University of Hong Kong · Nvidia AI Technology Center +2

AI Agents for Scientific DiscoveryMultimodal Contrastive LearningMultimodal AgentsDrug Discovery
AskPDF
Jun 2, 2026

Perceive Before Reasoning: A Pre-Reasoning Perception Framework for Efficient and Reliable Proactive Mobile Agents

Multimodal large language models (MLLMs) have substantially advanced mobile agents, yet proactive mobile assistance remains challenging because agents must decide when to intervene before determining how to assist. Existing systems often implement these two decisions within a unified MLLM-based pipeline, leading to goal misalignment between conservative intervention filtering and comprehensive assistance generation, as well as redundant inference when the agent should remain silent. To address these limitations, we propose the Pre-Reasoning Perception Framework (PRPF), a two-stage framework built on perceiving before reasoning. PRPF introduces a lightweight Multimodal Proactive Perceptor (MPP) for intervention gating and context compression, and activates the Proactive Agent Reasoner (PAR) only when intervention is warranted. Experiments on the ProactiveMobile benchmark show that PRPF substantially reduces false trigger rates (FTR) while improving success rates (SR) and inference efficiency over the ProactiveMobile baseline.

Zhijie Ding, Weinan Hong, Zicheng Zhu +6

HyperAI Team, Xiaomi Corporation · Zhongnan University of Economics and Law · 1HyperAI Team, Xiaomi Corporation +5

Efficient Multimodal InferenceProactive AssistanceMultimodal AgentsLLM Agent Reliability
AskPDF
Jun 2, 2026

PhotoCraft: Agentic Reasoning with Hierarchical Self-Evolving Memory for Deep Image Search

Deep Image Search requires multi-step reasoning over rich contextual cues, such as time, location, and event relations. However, most existing LLM-based agents are stateless and reactive, lacking persistent memory to maintain long-horizon context or transfer experience across tasks, which often leads to execution drift and experience isolation. To address these limitations, we propose PhotoCraft, a training-free, hierarchical memory system for photo-search agents. Inspired by human cognition, PhotoCraft equips MLLMs with working, episodic, and semantic memory, which are dynamically invoked during reasoning to preserve logical consistency and knowledge transferability throughout multi-step reasoning and answer generation. Extensive experiments on DISBench demonstrate that PhotoCraft consistently improves context-aware retrieval across diverse MLLM backbones, achieving gains of up to 18.5% and effectively mitigating key bottlenecks in memoryless deep image search, offering a practical path toward reliable and generalizable multimodal search agents.

Kailin Lyu, Zhiqiang Yuan, Jianwei He +9

Pattern Recognition Center, WeChat AI, Tencent Inc. · Institute of Automation, Chinese Academy of Sciences · Zhongguancun Academy +2

Multimodal Search AgentsHierarchical MemoryImage RetrievalMultimodal AgentsMemory-Augmented Language Model Agents
AskPDF
Jun 2, 2026

MUSE: A Unified Agentic Harness for MLLMs

Despite rapid progress, multimodal large language models (MLLMs) still fail on tasks that humans solve effortlessly, such as navigating a grid maze from a screenshot or selecting the correct puzzle piece. Rather than retraining the model, we ask a complementary question: how much capability can be elicited from a frozen MLLM purely by improving the execution scaffold around it? We introduce MUSE, a multimodal unified structured execution harness that wraps any off-the-shelf MLLM with composable modules for task representation, visual processing, perception tool use, structured parsing, deterministic verification, and verifier-guided repair, without any model retraining. We evaluate MUSE across diverse benchmarks spanning visual spatial planning, visual perception, multimodal reasoning, and fine-grained visual discrimination, using multiple state-of-the-art MLLMs. MUSE delivers consistent gains over the bare model in all settings, with the largest jumps on challenging instances. Further analysis reveals that many MLLM failures arise from harness-level shortcomings rather than fundamental model deficits, and can be addressed through verifier-guided repair without touching the model. These findings highlight the agentic multimodal harness as a critical yet underexplored design dimension, offering an orthogonal avenue for improving MLLMs beyond model-centric optimization.

Jianglin Lu, Hailing Wang, Xu Ma +4

Northeastern University

Multimodal ReasoningMultimodal AgentsLLM Agent HarnessesMultimodal Understanding
AskPDF
Jun 1, 2026

HLL: Can Agents Cross Humanity's Last Line of Verification?

Multimodal agents are increasingly expected to operate interfaces on behalf of users, raising a central deployment question: can they truly substitute for humans in workflows that services deliberately protect against automation? CAPTCHA verification makes this question concrete. It is not merely a visual puzzle, but a human-verification boundary placed before account creation, content access, form submission, and other protected actions. We introduce Humanity's Last Line of Verification (HLL), a controlled benchmark that uses interactive CAPTCHA verification to evaluate whether agents can cross this boundary through grounded, human-like interaction rather than recognition alone. HLL covers diverse CAPTCHA interactions and exposes agents to controlled realism stressors, including cluttered webpages, harder task variants, and trace-conditioned validation of the solving process. We evaluate eight frontier multimodal agents in a closed-loop GUI environment. The results show that current agents remain brittle at this human-substitution boundary: performance varies sharply across verification types, degrades under realistic interface conditions, and drops further when correct answers must be supported by valid action traces. By exposing gaps in localization, action calibration, state tracking, and process consistency, HLL provides a concrete testbed for measuring how close multimodal agents are to acting as human substitutes in protected real-world workflows. Our code is available at https://github.com/XinhaoS0101/HLL

Xinhao Song, Su Su, Sirui Song +6

1Shanghai Jiao Tong University · 2Shandong University · 3Tongji University

GUI AgentsAI Agent EvaluationAI Agent BenchmarksMultimodal Agents
AskPDF
Jun 1, 2026

TVIR: Building Deep Research Agents Towards Text-Visual Interleaved Report Generation

Deep Research Agents have shown strong capability in multi-step information retrieval, reasoning, and long-form report generation, but existing benchmarks and systems remain predominantly text-centric, with limited evaluation of whether visual elements are factually reliable and well aligned with the surrounding analysis. To address this gap, we introduce TVIR (Text-Visual Interleaved Report Generation), which includes TVIR-Bench, a benchmark of 100 expert-curated multimodal deep research tasks that require visual elements to serve specific analytical sub-goals, and TVIR-Agent, a hierarchical multi-agent framework that serves as a strong baseline for constructing outlines, retrieving images, generating charts with traceable sources, and composing reports through context-aware sequential writing. We further develop a dual-path evaluation framework that combines Textual Assessment and Visual Assessment. Experiments across nine deep research systems show that TVIR-Agent achieves strong overall performance, underscoring the importance of explicit multimodal design and evaluation for evidence-driven report generation.

Xinkai Ma, Zhiqi Bai, Dingling Zhang +21

Nanjing University · Alibaba Group

Deep Research AgentsAI Agent BenchmarksMultimodal AgentsMultimodal Model Evaluation
AskPDF
Jun 1, 2026

MetaForge: A Self-Evolving Multimodal Agent that Retrieves, Adapts, and Forges Tools On Demand

Multimodal agents have achieved notable progress on complex reasoning tasks through tool use, yet remain limited by two issues: statically predefined tool inventories fail to generalize to unseen scenarios, and indiscriminate tool invocation incurs redundant cost and noise-induced errors. We propose MetaForge, a multimodal agent framework that learns when to invoke tools and how to evolve its toolset on demand. MetaForge factorizes agentic behavior into four coupled stages: Decide (judging whether tool use is warranted), Retrieve (selecting suitable tools), Adapt (grounding tool parameters in task context), and Forge (synthesizing new skills online and recycling them into the tool library for reuse), forming a closed judge-retrieve-adapt-forge-recycle loop. A unified orchestration policy enables the agent to choose among answering directly, reusing existing tools, or forging new ones. We jointly optimize invocation necessity, retrieval accuracy, execution effectiveness, and forged-skill reusability via reinforcement learning, with an explicit invocation-cost penalty discouraging redundant calls. Across 12 benchmarks, MetaForge consistently surpasses 16 baselines in accuracy, efficiency, and generalization, validating a paradigm shift from static tool inventories to on-demand self-evolution.

Shouang Wei, Houcheng Min, Xinpeng Dong +8

East China Normal University · Zhejiang University · Tsinghua University +2

LLM Agent Skill LearningMultimodal AgentsSelf-Evolving AgentsRL for Tool UseMultimodal Tool Use
AskPDF
May 31, 2026

Crayotter: Traceable Multi-Agent Workflows for Long-Form Video Editing

Editing a long-form video from heterogeneous footage requires more than selecting clips: an agent must preserve narrative intent across material preparation, timeline construction, post-production, and revision while leaving enough evidence to diagnose failures. We present Crayotter, an open-source multimodal multi-agent system for prompt-driven video editing. Crayotter organizes production into three phases: coverage-aware material preparation, artifact-based editing research, and tool-grounded timeline execution. Each phase externalizes inspectable artifacts, including coverage reports, multimodal analyses, editing blueprints, tool calls, and intermediate renders. These artifacts make an editing run traceable and allow failed segments to be diagnosed and selectively revised instead of requiring a full restart. We evaluate Crayotter on 23 editing themes against CapCut-Mate and CutClaw. Under human evaluation, Crayotter achieves an average score of 3.40/5, compared with 2.44 and 1.70 for the two baselines, with consistent gains in theme alignment, narrative coherence, and editing smoothness. We additionally describe a replayable trajectory schema and verifiable reward design that prepare these workflows for future policy optimization. Code, traces, and examples are publicly available at https://github.com/idwts/Crayotter.

Lecheng Yan, Yichong Zhang, Ben Pan +5

Crayotter Project

Video EditingMulti-Agent CollaborationInstruction-Based Video EditingMultimodal AgentsMultimodal Tool Use
AskPDF
May 30, 2026

Sandboxed Coding Agents are Competitive Omni-modal Task Solvers

As multimodal LLMs increasingly target video and audio, it is often assumed that such tasks require native omnimodal models. We show that this is not always the case: coding agents with only text+image access and a sandboxed tool-use interface can match, and in several settings outperform, SOTA native omnimodal models and predefined multimodal agent scaffolds across multiple audio-video benchmarks. Our trajectory analysis suggests that their strength comes from writing code and orchestrating tools to extract relevant evidence from transcripts, frames, and other modality signals, thereby converting omnimodal tasks into retrieval and information-processing problems rather than ingesting entire media streams. We further characterize their limitations through a failure taxonomy and process-level trace analysis, and show that simple skill injection, including human-written and self-distilled skills, substantially improves performance. To explore open-source elicitation, we introduce Code-X, a training recipe with the OmniCoding trajectory dataset and verifiable reward, and provide baselines on Qwen-3.5-9B and Qwen-3.6-27B. Finally, we argue that the next frontier is many-modality processing, and introduce TerminalBench-O, a process-level benchmark for real-world omnimodal processing tasks. Code will be available at https://github.com/Dongping-Chen/OmniCoding.

Dongping Chen, Xuanao Huang, Zhihan Hu +3

University of Maryland · MBZUAI

Audio-Visual UnderstandingAI Coding AgentsAI Agent BenchmarksMultimodal AgentsMultimodal UnderstandingMultimodal Tool Use
AskPDF
May 29, 2026

Task-Focused Memorization for Multimodal Agents

Long-term memory is essential for multimodal agents to build coherent experience, accumulate world knowledge, and achieve continual learning. However, constructing effective memory goes beyond memory module design and basic requirements such as accuracy and fidelity; the key challenge lies in determining what to memorize. Multimodal agents, such as embodied agents, continuously perceive, reason, and act in real or virtual environments, receiving an unbounded stream of multimodal observations. From this combinatorial explosion of information, an agent must selectively retain content that is relevant to its role in the environment and valuable for future tasks. To bridge this gap, we frame memory generation as a learnable memorization policy and introduce TaskMem (Task-focused Memorization Policy Learning), a reinforcement-learning-based framework that enables the policy to dynamically adjust its focus to the demands of real tasks encountered in the environment. TaskMem adopts a two-phase training paradigm: Phase One learns how to memorize by optimizing memory quality under fundamental fidelity requirements; Phase Two occurs after deployment, where the agent learns what to memorize by tuning an adapter on its base MLLM, using recent environment tasks to define a reward model that guides the memorization policy toward task-relevant content. To evaluate our approach, we reformulate VideoMME, EgoLife, and EgoTempo into streaming benchmarks that simulate a realistic setting in which an agent processes streaming observations and handles tasks arriving online. To isolate memory assessment, the questions must be answered using only the agent's memory, without access to raw video. Built on Qwen3-VL-30B-A3B, TaskMem improves VQA accuracy by 6.3%, 7.0%, and 5.3% on these benchmarks, respectively.

Tao Zou, Yichen He, Tian Qiu +2

ByteDance Seed · Fudan University

Continual Learning for LLM AgentsMultimodal MemoryMultimodal AgentsRL for LLM Agents
AskPDF
May 29, 2026

MineExplorer: Evaluating Open-World Exploration of MLLM Agents in Minecraft

Multimodal large language models (MLLMs) have shown strong capabilities in perception, reasoning, and action generation. However, their ability to sustain exploration in dynamic open worlds remains unclear. Existing embodied and game-based benchmarks often compress interaction into short-horizon tasks or entangle success with domain-specific game mechanics. In this paper, we introduce MineExplorer benchmark for evaluating open-world exploration capabilities of MLLM agents in Minecraft. We first filter atomic tasks whose solutions rely heavily on Minecraft-specific knowledge to better reflect general open-world reasoning. Then we organize the benchmark around a ReAct-style capability formulation and compose atomic tasks into implicit multi-hop tasks. To further construct reliable instances, MineExplorer uses a multi-agent synthesis workflow that jointly designs task graphs, sandbox scenes, and rule-based milestone evaluators. Human evaluation shows that the multi-agent synthesis workflow produces significantly more reliable instances than a single-agent baseline. Experiments with advanced MLLM agents show that open-world exploration remains challenging, as strong models can handle many single-hop tasks but degrade sharply when hidden prerequisites must be coordinated over longer trajectories. Further analysis finds that task difficulty tracks agent completion, and larger models or thinking modes do not consistently translate into better performance. Code and dataset are available at https://github.com/Jometeorie/MineExplorer.

Tianjie Ju, Yueqing Sun, Zheng Wu +7

School of Computer Science, Shanghai Jiao Tong University · 2Meituan

Long-Horizon Agent EvaluationAI Agent BenchmarksMultimodal Agents
AskPDF
May 28, 2026

A Theory-Guided LLM Pedagogical Agent for STEM+C Scaffolding Without Over-Reliance

LLM pedagogical agents are proliferating, yet recent findings have raised questions about their adherence to established theories of learning and, by extension, their educational value. Concerns regarding cognitive offloading, over-reliance, and "gaming" behaviors persist and remain largely unaddressed. In response, we developed Copa, an agentic, multi-agent, multimodal Collaborative Peer Agent for STEM+C learning. Copa is built on top of the Evidence-Decision-Feedback (EDF) framework, grounding its interactions in Social Cognitive Theory and Social Constructivism and promoting sense-making through adaptive, dialogic support rather than answer-seeking. In an authentic high school computational-modeling study (n=33 dyads), we demonstrate that Copa (1) supports students' confidence building and ability to verbalize conceptual understanding without causing dependence; and (2) provides adaptive feedback personalized to learners that is interpretable with respect to students' multimodal input data. These findings position theory-guided, multimodal LLM agents as a promising path toward classroom AI integration that amplifies students' reasoning rather than replacing it.

Clayton Cohn, Surya Rayala, Siyuan Guo +13

College of Connected Computing, Vanderbilt University, Nashville, Tennessee 37240 · The School for Science and Math at Vanderbilt, Vanderbilt University, Nashville, Tennessee 37240

Multi-Agent LLM SystemsAI in EducationMultimodal AgentsCognitive Offloading
AskPDF
May 28, 2026

VideoFDB: Evaluating Full-Duplex Vision-Speech Capabilities in Conversational Agents

Natural human conversation is full-duplex and audio-visual: people simultaneously speak and listen while continuously interpreting and producing nonverbal cues, such as nods, smiles, and gestures. To support successful human-agent interaction, agents must model full-duplex audiovisual conversation; however, existing full-duplex benchmarks evaluate only speech. In this work, we present VideoFDB, the first benchmark to evaluate full-duplex audio-visual-to-audio-visual (AV2AV) conversational agents. VideoFDB contributes (i) 237 dyadic clips spanning 11 nonverbal conversational dynamics from real-world video calls, (ii) a taxonomy separating perception from generation behaviors, and (iii) a rubric-based LM-as-judge evaluation framework with interpretable axes for assessing conversational quality with respect to nonverbal conversational dynamics. Across open- and closed-source vision-speech agents, we find systematic failure modes: captioning collapse and visual-stream ignorance, and we show that current systems exploit vision for explicit visual question answering but not for the streaming joint audiovisual grounding required in natural conversation. We further evaluate cascaded speech-to-avatar systems and find that their architecture fundamentally precludes the production of full-duplex nonverbal cues. As the first benchmark for full-duplex AV2AV interaction, VideoFDB establishes a foundation for systematic evaluation and, we hope, will accelerate the advancement and development of next-generation multimodal conversational agents.

Amrita Mazumdar, Seonwook Park, Rajarshi Roy +6

NVIDIA · David AI

Multimodal AgentsMultimodal Model EvaluationFull-Duplex Spoken Dialogue Systems
AskPDF
May 28, 2026

Towards Verifiable Multimodal Deep Research: A Multi-Agent Harness for Interleaved Report Generation

Large Language Models (LLMs) have advanced autonomous agents from deep search, which retrieves concise factual answers, to deep research, which synthesizes scattered evidence into long-form reports. However, verifiable multimodal deep research remains challenging due to open-ended synthesis without deterministic ground truth and the need to interleave textual arguments with visual evidence. We propose Ptah, a multi-agent harness for interleaved report generation. Ptah orchestrates the lifecycle from user query to rendered web report through planning, research, and writing stages, where specialized agents construct visual-aware plans, collect claim-grounded evidence, maintain source-aligned images in a Visual Working Memory, and compose reports through declarative multimodal tool use. A verifier agent serves as the harness's acceptance function, enforcing factual grounding, citation fidelity, and cross-modal consistency throughout the workflow. We further introduce PtahEval, an evaluation protocol that augments existing benchmarks with image-level and presentation-level assessments. Experiments on deep research benchmarks show that Ptah produces more reliable, visually informative, and usable human-facing multimodal reports than strong baselines. Our code is released at https://github.com/SnowNation101/Ptah

Chenghao Zhang, Guanting Dong, Yufan Liu +3

Gaoling School of Artificial Intelligence, Renmin University of China

Multimodal GenerationMultimodal AgentsMulti-Agent LLMsMultimodal Model EvaluationMultimodal Tool Use
AskPDF
May 28, 2026

AgentCVR: Active Multi-Agent Cross-Video Reasoning via Script-Simulated Reinforcement Learning

Cross-Video Reasoning (CVR) has emerged as a critical frontier in multimodal intelligence, requiring models to retrieve, align, and aggregate evidence distributed across multiple videos. Current Multimodal Large Language Models (MLLMs) often struggle with CVR, as simple single-pass strategies encode multiple videos into a shared compressed context, potentially obscuring rare but critical evidence. In this paper, we propose AgentCVR, a multi-agent framework that treats CVR as an active evidence-acquisition task. AgentCVR employs a Master Agent to iteratively coordinate specialized Visual and Audio Agents for targeted evidence extraction. To ensure efficient training, we introduce Script-Simulated RL, which optimizes the agent's policy with LLM-generated semantic scripts and a lightweight text-based simulator, bypassing costly multimodal inference during online exploration. Experimental results on a comprehensive CVR benchmark show that AgentCVR outperforms single-pass baselines and achieves comparable performance to state-of-the-art closed-source systems, particularly in complex cross-video alignment and localization. To ensure reproducibility, our code is available at https://github.com/wang-jh24/AgentCVR.

Yilun Qiu, Jiahe Wang, Cilin Yan +4

Xiaohongshu Inc. · Tsinghua Shenzhen International Graduate School, Tsinghua University

Video ReasoningMultimodal AgentsRL for VLMs
AskPDF
May 27, 2026

VFEAgent: A Multimodal Agent Framework for End-to-End Automated Finite Element Analysis

Finite Element Analysis (FEA) serves as the cornerstone of modern engineering design. However, its workflow is inherently complex and relies heavily on domain expertise. Although recent efforts have integrated Large Language Models (LLMs) into FEA, existing approaches face limitations in handling multimodal inputs and executing complex tasks. To address these limitations, we propose VFEAgent, an end-to-end multi-agent system designed to automate FEA modeling and simulation directly from input images and problem descriptions. Our methodology integrates two core components: (1) a multimodal vision-language multi-agent pipeline that employs ReAct-driven reasoning to extract structured FEA specifications from heterogeneous inputs and (2) a verification-first code synthesis framework, incorporating robust self-debugging and fallback mechanisms to ensure executability and physical validity. We systematically evaluated the system across various engineering mechanics scenarios. The results demonstrate that VFEAgent achieves a high success rate in generating complete and physically valid simulations, outperforming LLM-based baseline methods in reliability and correctness. These findings validate the feasibility of automating the complete FEA workflow, highlighting the framework's potential to liberate engineers from tedious manual analysis.

Jiachen Zhang, Junyi Lao, Chenghao Liu +5

Peking University · China Agricultural University · 1Peking University +1

Program SynthesisFinite Element MethodMultimodal AgentsMulti-Agent LLMs
AskPDF