Multimodal Reasoning

Recent progress in multimodal large language models has led to strong performance on reasoning tasks, but these improvements largely rely on high-quality annotated data or teacher-model distillation, both of which are costly and difficult to scale. To address this, we propose an unsupervised self-evolution training framework for multimodal reasoning that achieves stable performance improvements without using human-annotated answers or external reward models. For each input, we sample multiple reasoning trajectories and jointly model their within group structure. We use the Actor's self-consistency signal as a training prior, and introduce a bounded Judge based modulation to continuously reweight trajectories of different quality. We further model the modulated scores as a group level distribution and convert absolute scores into relative advantages within each group, enabling more robust policy updates. Trained with Group Relative Policy Optimization (GRPO) on unlabeled data, our method consistently improves reasoning performance and generalization on five mathematical reasoning benchmarks, offering a scalable path toward self-evolving multimodal models. The code are available at https://github.com/OPPO-Mente-Lab/LLM-Self-Judge.

Composed Video Retrieval (CoVR) aims to find a target video given a reference video and a textual modification. Prior work assumes the modification text fully specifies the visual changes, overlooking after-effects and implicit consequences (e.g., motion, state transitions, viewpoint or duration cues) that emerge from the edit. We argue that successful CoVR requires reasoning about these after-effects. We introduce a reasoning-first, zero-shot approach that leverages large multimodal models to (i) infer causal and temporal consequences implied by the edit, and (ii) align the resulting reasoned queries to candidate videos without task-specific finetuning. To evaluate reasoning in CoVR, we also propose CoVR-Reason, a benchmark that pairs each (reference, edit, target) triplet with structured internal reasoning traces and challenging distractors that require predicting after-effects rather than keyword matching. Experiments show that our zero-shot method outperforms strong retrieval baselines on recall at K and particularly excels on implicit-effect subsets. Our automatic and human analysis confirm higher step consistency and effect factuality in our retrieved results. Our findings show that incorporating reasoning into general-purpose multimodal models enables effective CoVR by explicitly accounting for causal and temporal after-effects. This reduces dependence on task-specific supervision, improves generalization to challenging implicit-effect cases, and enhances interpretability of retrieval outcomes. These results point toward a scalable and principled framework for explainable video search. The model, code, and benchmark are available at https://github.com/mbzuai-oryx/CoVR-R.

Rapid identification of hazardous events is essential for next-generation Earth Observation (EO) missions supporting disaster response. However, current monitoring pipelines remain largely ground-centric, introducing latency due to downlink limitations, multi-source data fusion constraints, and the computational cost of exhaustive scene analysis. This work proposes a hierarchical multi-agent architecture for onboard EO processing under strict resource and bandwidth constraints. The system enables the exploitation of complementary multimodal observations by coordinating specialized AI agents within an event-driven decision pipeline. AI agents can be deployed across multiple nodes in a distributed setting, such as satellite platforms. An Early Warning agent generates fast hypotheses from onboard observations and selectively activates domain-specific analysis agents, while a Decision agent consolidates the evidence to issue a final alert. The architecture combines vision-language models, traditional remote sensing analysis tools, and role-specialized agents to enable structured reasoning over multimodal observations while minimizing unnecessary computation. A proof-of-concept implementation was executed on the engineering model of an edge-computing platform currently deployed in orbit, using representative satellite data. Experiments on wildfire and flood monitoring scenarios show that the proposed routing-based pipeline significantly reduces computational overhead while maintaining coherent decision outputs, demonstrating the feasibility of distributed agent-based reasoning for future autonomous EO constellations.

Multimodal large language models (MLLMs) perform strongly on natural images, yet their ability to understand discrete visual symbols remains unclear. We present a multi-domain benchmark spanning language, culture, mathematics, physics and chemistry, organized into three cognitive levels: perception and recognition, combination and reasoning, and association and critical thinking. Across leading MLLMs, we observe a consistent cognitive mismatch. Models frequently underperform on elementary symbol recognition while appearing relatively competent on more complex reasoning tasks. This recognition-reasoning inversion indicates that current systems often compensate with linguistic priors, template retrieval or procedural reasoning instead of robust visual grounding. The pattern is especially clear for sparse, low-redundancy symbols such as handwritten characters, formula graphs, circuit diagrams and chemical structures. These results show that symbolic understanding remains a major bottleneck for multimodal intelligence and motivate training and evaluation schemes that prioritize grounded perception in discrete semantic spaces.

The widespread adoption of dashcams has made video evidence in traffic accidents increasingly abundant, yet transforming "what happened in the video" into "who is responsible under which legal provisions" still relies heavily on human experts. Existing ego-view traffic accident studies mainly focus on perception and semantic understanding, while LLM-based legal methods are mostly built on textual case descriptions and rarely incorporate video evidence, leaving a clear gap between the two. We first propose C-TRAIL, a multimodal legal dataset that, under the Chinese traffic regulation system, explicitly aligns dashcam videos and textual descriptions with a closed set of responsibility modes and their corresponding Chinese traffic statutes. On this basis, we introduce a two-stage framework: (1) a traffic accident understanding module that generates textual video descriptions; and (2) a legal multi-agent framework that outputs responsibility modes, statute sets, and complete judgment reports. Experimental results on C-TRAIL and MM-AU show that our method outperforms general and legal LLMs, as well as existing agent-based approaches, while providing a transparent and interpretable legal reasoning process.

Dental triage is a safety-critical clinical routing task that requires integrating multimodal clinical information (e.g., patient complaints and radiographic evidence) to determine complete referral plans. We present Dental-TriageBench, the first expert-annotated benchmark for reasoning-driven multimodal dental triage. Built from authentic outpatient workflows, it contains 246 de-identified cases annotated with expert-authored golden reasoning trajectories, together with hierarchical triage labels. We benchmark 19 proprietary, open-source, and medical-domain MLLMs against three junior dentists serving as the human baseline, and find a substantial human--model gap, on fine-grained treatment-level triage. Further analyses show that accurate triage requires both complaint and OPG information, and that model errors concentrate on cases with multiple referral domains, where MLLMs tend to produce overly narrow referral sets and omission-heavy errors. Dental-TriageBench provides a realistic testbed for developing multimodal clinical AI systems that are more clinically grounded, coverage-aware, and safer for downstream care.

The efficient adjudication of responsibility disputes is pivotal for maintaining marketplace fairness. However, the exponential surge in ride-hailing volume renders manual review intractable, while conventional automated methods lack the reasoning transparency required for quasi-judicial decisions. Although Multimodal LLMs offer a promising paradigm, they fundamentally struggle to bridge the gap between general visual semantics and rigorous evidentiary protocols, often leading to perceptual hallucinations and logical looseness. To address these systemic misalignments, we introduce RideJudge, a Progressive Visual-Logic-Aligned Framework. Instead of relying on generic pre-training, we bridge the semantic gap via SynTraj, a synthesis engine that grounds abstract liability concepts into concrete trajectory patterns. To resolve the conflict between massive regulation volume and limited context windows, we propose an Adaptive Context Optimization strategy that distills expert knowledge, coupled with a Chain-of-Adjudication mechanism to enforce active evidentiary inquiry. Furthermore, addressing the inadequacy of sparse binary feedback for complex liability assessment, we implement a novel Ordinal-Sensitive Reinforcement Learning mechanism that calibrates decision boundaries against hierarchical severity. Extensive experiments show that our RideJudge-8B achieves 88.41% accuracy, surpassing 32B-scale baselines and establishing a new standard for interpretable adjudication.

Surgical intelligence has the potential to improve the safety and consistency of surgical care, yet most existing surgical AI frameworks remain task-specific and struggle to generalize across procedures and institutions. Although multimodal foundation models, particularly multimodal large language models, have demonstrated strong cross-task capabilities across various medical domains, their advancement in surgery remains constrained by the lack of large-scale, systematically curated multimodal data. To address this challenge, we introduce Surg$Σ$, a spectrum of large-scale multimodal data and foundation models for surgical intelligence. At the core of this framework lies Surg$Σ$-DB, a large-scale multimodal data foundation designed to support diverse surgical tasks. Surg$Σ$-DB consolidates heterogeneous surgical data sources (including open-source datasets, curated in-house clinical collections and web-source data) into a unified schema, aiming to improve label consistency and data standardization across heterogeneous datasets. Surg$Σ$-DB spans 6 clinical specialties and diverse surgical types, providing rich image- and video-level annotations across 18 practical surgical tasks covering understanding, reasoning, planning, and generation, at an unprecedented scale (over 5.98M conversations). Beyond conventional multimodal conversations, Surg$Σ$-DB incorporates hierarchical reasoning annotations, providing richer semantic cues to support deeper contextual understanding in complex surgical scenarios. We further provide empirical evidence through recently developed surgical foundation models built upon Surg$Σ$-DB, illustrating the practical benefits of large-scale multimodal annotations, unified semantic design, and structured reasoning annotations for improving cross-task generalization and interpretability.

Open-Vocabulary Multimodal Emotion Recognition (OV-MER) is inherently challenging due to the ambiguity of equivocal multimodal cues, which often stem from distinct unobserved situational dynamics. While Multimodal Large Language Models (MLLMs) offer extensive semantic coverage, their performance is often bottlenecked by premature commitment to dominant data priors, resulting in suboptimal heuristics that overlook crucial, complementary affective cues across modalities. We argue that effective affective reasoning requires more than surface-level association; it necessitates reconstructing nuanced emotional states by synthesizing multiple evidence-grounded rationales that reconcile these observations from diverse latent perspectives. We introduce HyDRA, a Hybrid-evidential Deductive Reasoning Architecture that formalizes inference as a Propose-Verify-Decide protocol. To internalize this abductive process, we employ reinforcement learning with hierarchical reward shaping, aligning the reasoning trajectories with final task performance to ensure they best reconcile the observed multimodal cues. Systematic evaluations validate our design choices, with HyDRA consistently outperforming strong baselines--especially in ambiguous or conflicting scenarios--while providing interpretable, diagnostic evidence traces.

The rapid advancement of Multimodal Large Language Models (MLLMs) has enabled browsing agents to acquire and reason over multimodal information in the real world. But existing benchmarks suffer from two limitations: insufficient evaluation of visual reasoning ability and the neglect of native visual information of web pages in the reasoning chains. To address these challenges, we introduce a new benchmark for visual-native search, VisBrowse-Bench. It contains 169 VQA instances covering multiple domains and evaluates the models' visual reasoning capabilities during the search process through multimodal evidence cross-validation via text-image retrieval and joint reasoning. These data were constructed by human experts using a multi-stage pipeline and underwent rigorous manual verification. We additionally propose an agent workflow that can effectively drive the browsing agent to actively collect and reason over visual information during the search process. We comprehensively evaluated both open-source and closed-source models in this workflow. Experimental results show that even the best-performing model, Claude-4.6-Opus only achieves an accuracy of 47.6%, while the proprietary Deep Research model, o3-deep-research only achieves an accuracy of 41.1%. The code and data can be accessed at: https://github.com/ZhengboZhang/VisBrowse-Bench

Existing instruction-based image editing models perform well with simple, single-step instructions but degrade in realistic scenarios that involve multiple, lengthy, and interdependent directives. A main cause is the scarcity of training data with complex multi-instruction annotations. However, it is costly to collect such data and retrain these models. To address this challenge, we propose MSRAMIE, a training-free agent framework built on Multimodal Large Language Model (MLLM). MSRAMIE takes existing editing models as plug-in components and handle multi-instruction tasks via structured multimodal reasoning. It orchestrates iterative interactions between an MLLM-based Instructor and an image editing Actor, introducing a novel reasoning topology that comprises the proposed Tree-of-States and Graph-of-References. During inference, complex instructions are decomposed into multiple editing steps which enable state transitions, cross-step information aggregation, and original input recall, which enables systematic exploration of the image editing space and flexible progressive output refinement. The visualizable inference topology further provides interpretable and controllable decision pathways. Experiments show that as the instruction complexity increases, MSRAMIE can improve instruction following over 15% and increases the probability of finishing all modifications in a single run over 100%, while preserving perceptual quality and maintaining visual consistency.

Large Language Models (LLMs) have demonstrated exceptional logical reasoning capabilities but frequently struggle with the continuous spatiotemporal dynamics governed by Partial Differential Equations (PDEs), often resulting in non-physical hallucinations. Existing approaches typically resort to costly, domain-specific fine-tuning, which severely limits cross-domain generalization and interpretability. To bridge this gap, we propose OMNIFLOW, a neuro-symbolic architecture designed to ground frozen multimodal LLMs in fundamental physical laws without requiring domain-specific parameter updates. OMNIFLOW introduces a novel Semantic-Symbolic Alignment mechanism that projects high-dimensional flow tensors into topological linguistic descriptors, enabling the model to perceive physical structures rather than raw pixel values. Furthermore, we construct a Physics-Guided Chain-of-Thought (PG-CoT) workflow that orchestrates reasoning through dynamic constraint injection (e.g., mass conservation) and iterative reflexive verification. We evaluate OMNIFLOW on a comprehensive benchmark spanning microscopic turbulence, theoretical Navier-Stokes equations, and macroscopic global weather forecasting. Empirical results demonstrate that OMNIFLOW significantly outperforms traditional deep learning baselines in zero-shot generalization and few-shot adaptation tasks. Crucially, it offers transparent, physically consistent reasoning reports, marking a paradigm shift from black-box fitting to interpretable scientific reasoning.

Multimodal Large Language Models (MLLMs) have recently made rapid progress toward unified Omni models that integrate vision, language, and audio. However, existing environments largely focus on 2D or 3D visual context and vision-language tasks, offering limited support for temporally dependent auditory signals and selective cross-modal integration, where different modalities may provide complementary or interfering information, which are essential capabilities for realistic multimodal reasoning. As a result, whether models can actively coordinate modalities and reason under time-varying, irreversible conditions remains underexplored. To this end, we introduce EscapeCraft-4D, a customizable 4D environment for assessing selective cross-modal perception and time awareness in Omni models. It incorporates trigger-based auditory sources, temporally transient evidence, and location-dependent cues, requiring agents to perform spatio-temporal reasoning and proactive multimodal integration under time constraints. Building on this environment, we curate a benchmark to evaluate corresponding abilities across powerful models. Evaluation results suggest that models struggle with modality bias, and reveal significant gaps in current model's ability to integrate multiple modalities under time constraints. Further in-depth analysis uncovers how multiple modalities interact and jointly influence model decisions in complex multimodal reasoning environments.

Recent advances in Multimodal Large Language Models (MLLMs) have shown impressive reasoning capabilities across vision-language tasks, yet still face the challenge of compute-difficulty mismatch. Through empirical analyses, we identify that existing decoding methods may waste compute on easy cases while underserving hard ones, affecting both model effectiveness and efficiency. To address this issue, we first develop a theoretical framework that links sampling coverage, instance difficulty, and residual risk. Our analysis reveals that multimodal reasoning exhibits a heavy-tailed difficulty distribution; a small subset of hard or ambiguous samples dominates the residual failure probability. Based on this insight, we propose Coverage-Aware Multimodal Decoding (CAMD), an adaptive inference mechanism that dynamically allocates computation according to estimated uncertainty. CAMD integrates evidence-weighted scoring, posterior coverage estimation, and sequential Bayesian updating to balance efficiency and reliability under a limited token budget. Experiments on various benchmark datasets and baselines demonstrate the effectiveness and advantages of our approach.

Modern agricultural data is sourced from diverse platforms and spans multiple spatial scales, ranging from ground-level close-up photography to Unmanned Aerial Vehicle (UAV) aerial observation and satellite remote sensing imagery. Accordingly, agricultural multimodal reasoning demands robust cross-scale spatial understanding. However, due to the lack of multi-view agricultural benchmark datasets, existing multimodal large language models (MLLMs) exhibit severe ground-level bias, which leads to scale confusion then semantic collapse in agricultural perception tasks, such as misinterpreting farmland imagery as walls or floors. To address this, we introduce AgroOmni, a large-scale multi-view training corpus with 288K Visual Question Answering pairs covering 56 specialized task categories across 14 task types, designed to capture diverse scales in modern precision agriculture. Built on this dataset, we propose AgroNVILA, which achieves a new state-of-the-art of 62.32% on the AgroMind benchmark (+15.03% over GPT-5.2), effectively mitigating the multi-view cross-scale gap for holistic agricultural understanding. Diagnostic evaluations on AgMMU further reveal an inherent heterogeneity between macro-priors and micro-diagnostics through constrained zero-shot performance. Meanwhile, even minimal fine-tuning leads to a dramatic performance gain of AgroNVILA on AgMMU, strongly demonstrating its generalization capability empowered by AgroOmni. Full training scripts are publicly available at https://anonymous.4open.science/r/AgroOmni-6510.

Multimodal Large Language Models (MLLMs) have shown strong performance in visual and audio understanding when evaluated in isolation. However, their ability to jointly reason over omni-modal (visual, audio, and textual) signals in long and complex videos remains largely unexplored. We introduce MMOU, a new benchmark designed to systematically evaluate multimodal understanding and reasoning under these challenging, real-world conditions. MMOU consists of 20,000 carefully curated questions paired with 11877 web-collected videos of varying length, spanning diverse domains and exhibiting rich, tightly coupled audio-visual content. The benchmark covers 13 fundamental skill categories, all of which require integrating evidence across modalities and time. All questions are manually annotated across multiple turns by professional annotators, ensuring high quality and reasoning fidelity. We evaluate 20+ state-of-the-art open-source and proprietary multimodal models on MMOU. The results expose substantial performance gaps: the best closed-source model achieves only 64.2% accuracy, while the strongest open-source model reaches just 46.8%. Our results highlight the challenges of long-form omni-modal understanding, revealing that current models frequently fail to apply even fundamental skills in long videos. Through detailed analysis, we further identify systematic failure modes and provide insights into where and why current models break.

Category-level multi-object trackers estimate trajectories for all detected instances, whereas some applications require propagating only targets specified by a user. We study offline, reference-grounded video tracking: given an ordered video clip, a natural-language query, and the reference-frame boxes of the specified targets, a model predicts their subsequent trajectories. We introduce QTrack, a vision-language policy trained with structured trajectory rewards and Temporal Perception-Aware Policy Optimization (TAPO). TAPO contrasts the likelihood of a generated response under the observed clip and a reference-frame repetition, encouraging the output to depend on visual evidence beyond the reference frame. We also introduce RMOT26, a controlled diagnostic benchmark constructed from existing tracking datasets. Under the reported protocol, QTrack obtains a motion- consistency score of 0.30, MOTP of 0.75, center-location error of 44.61 pixels, and normalized distance error of 0.39; its motion-consistency score ties the strongest comparison method. Component ablations suggest complementary effects from the motion reward and temporal contrast, while an exploratory four- clip, 50-frame evaluation examines transfer beyond the primary short-clip setting. The results support further study of temporal-evidence learning in trajectory- generating vision-language models. Code and data are available at https://github.com/gaash-lab/QTrack

In different multimodal scenarios, it needs to integrate and utilize information across modalities in a specific way based on the demands of the task. Different integration ways between modalities are referred to as "multimodal interaction". How well a model handles various multimodal interactions largely characterizes its multimodal ability. In this paper, we introduce MIBench, a comprehensive benchmark designed to evaluate the multimodal interaction capabilities of Large Multimodal Models (LMMs), which formulates each instance as a (con_v , con_t, task) triplet with contexts from vision and text, necessitating that LMMs employ correct forms of multimodal interaction to effectively complete the task. MIBench assesses models from three key aspects: the ability to source information from vision-centric or text-centric cues, and the ability to generate new information from their joint synergy. Each interaction capability is evaluated hierarchically across three cognitive levels: Recognition, Understanding, and Reasoning. MIBench comprises over 10,000 vision-text context pairs spanning 32 distinct tasks. Evaluation of state-of-the-art LMMs show that: (1) LMMs' ability on multimodal interaction remains constrained, despite the scaling of model parameters and training data; (2) they are easily distracted by textual modalities when processing vision information; (3) they mostly possess a basic capacity for multimodal synergy; and (4) natively trained multimodal models show noticeable deficits in fundamental interaction ability. We expect that these observations can serve as a reference for developing LMMs with more enhanced multimodal ability in the future.

Unified multimodal models target joint understanding, reasoning, and generation, but current image editing benchmarks are largely confined to natural images and shallow commonsense reasoning, offering limited assessment of this capability under structured, domain-specific constraints. In this work, we introduce GRADE, the first benchmark to assess discipline-informed knowledge and reasoning in image editing. GRADE comprises 520 carefully curated samples across 10 academic domains, spanning from natural science to social science. To support rigorous evaluation, we propose a multi-dimensional evaluation protocol that jointly assesses Discipline Reasoning, Visual Consistency, and Logical Readability. Extensive experiments on 20 state-of-the-art open-source and closed-source models reveal substantial limitations in current models under implicit, knowledge-intensive editing settings, leading to large performance gaps. Beyond quantitative scores, we conduct rigorous analyses and ablations to expose model shortcomings and identify the constraints within disciplinary editing. Together, GRADE pinpoints key directions for the future development of unified multimodal models, advancing the research on discipline-informed image editing and reasoning. Our benchmark and evaluation code are publicly released.

Constructing scientific multimodal document reasoning datasets for foundation model training involves an inherent trade-off among scale, faithfulness, and realism. To address this challenge, we introduce the synthesize-and-reground framework, a two-stage pipeline comprising: (1) Claim-Centric QA Synthesis, which generates faithful, isolated QA pairs and reasoning on focused segments, and (2) Document-Scale Regrounding, which programmatically re-embeds these pairs into full-document tasks to ensure realistic complexity. Using this framework, we construct SciMDR, a large-scale training dataset for cross-modal comprehension, comprising 300K QA pairs with explicit reasoning chains across 20K scientific papers. We further construct SciMDR-Eval, an expert-annotated benchmark to evaluate multimodal comprehension within full-length scientific workflows. Experiments demonstrate that models fine-tuned on SciMDR achieve significant improvements across multiple scientific QA benchmarks, particularly in those tasks requiring complex document-level reasoning.

Despite recent advances in multimodal reasoning, representing auxiliary geometric constructions remains a fundamental challenge for multimodal large language models (MLLMs). Such constructions are absent from the original diagram and must be introduced before theorems apply. Existing approaches predominantly rely on explicit construction paradigms, including text-based geometric specification, visual-token interleaving during reasoning, and tool-augmented geometric execution. However, these methods either fail to faithfully represent complex spatial relationships, incur representation mismatch between discrete symbols and continuous geometric structures, or rely on external capabilities that hinder end-to-end optimization. To address these limitations, we propose LatentGeo, a framework that learns continuous latent visual representations to internalize auxiliary geometric constructions without pixel-level rendering or external executors. We design a three-stage curriculum that progressively aligns and internalizes these latent representations through auxiliary visual supervision, followed by LaGDPO, a latent-aware reinforcement learning procedure that stabilizes latent representations during policy optimization while improving end-task correctness. To systematically evaluate construction-centric representation quality, we introduce GeoAux, a new benchmark targeting visually dependent geometry problems, and conduct experiments on GeoAux and MathVerse. Results show that LatentGeo achieves substantial gains on geometric reasoning tasks, particularly those requiring auxiliary constructions. Extensive analyses and ablation studies further validate the effectiveness of each component in our framework.

Existing chemical understanding tasks primarily rely on static molecular representations, limiting their ability to model inherently dynamic phenomena such as bond breaking or conformational changes, which are essential for a chemist to understand chemical reactions. To address this gap, we introduce Chemical Dynamics Understanding (ChemDU), a new task that translates 4D molecular trajectories into interpretable natural-language explanations. ChemDU focuses on fundamental dynamic scenarios, including gas-phase and catalytic reactions, and requires models to reason about key events along molecular trajectories, such as bond formation and dissociation, and to generate coherent, mechanistically grounded narratives. To benchmark this capability, we construct Chem4DBench, the first dataset pairing 4D molecular trajectories with expert-authored explanations across these settings. We further propose Chem4DLLM, a unified model that integrates an equivariant graph encoder with a pretrained large language model to explicitly capture molecular geometry and rotational dynamics. We hope that ChemDU, together with Chem4DBench and Chem4DLLM, will stimulate further research in dynamic chemical understanding and multimodal scientific reasoning.

Advancing towards artificial superintelligence requires rich and intelligent perceptual capabilities. A critical frontier in this pursuit is overcoming the limited spatial understanding of Multimodal Large Language Models (MLLMs), where geometry information is essential. Existing methods often address this by rigidly injecting geometric signals into every input, while ignoring their necessity and adding computation overhead. Contrary to this paradigm, our framework endows the model with an awareness of perceptual insufficiency, empowering it to autonomously engage geometric features in reasoning when 2D cues are deemed insufficient. To achieve this, we first introduce an independent geometry input channel to the model architecture and conduct alignment training, enabling the effective utilization of geometric features. Subsequently, to endow the model with perceptual awareness, we curate a dedicated spatial-aware supervised fine-tuning dataset. This serves to activate the model's latent internal cues, empowering it to autonomously determine the necessity of geometric information. Experiments across multiple spatial reasoning benchmarks validate this approach, demonstrating significant spatial gains without compromising 2D visual reasoning capabilities, offering a path toward more robust, efficient and self-aware multi-modal intelligence.

Unified multimodal models (UMMs) that integrate understanding, reasoning, generation, and editing face inherent trade-offs between maintaining strong semantic comprehension and acquiring powerful generation capabilities. In this report, we present InternVL-U, a lightweight 4B-parameter UMM that democratizes these capabilities within a unified framework. Guided by the principles of unified contextual modeling and modality-specific modular design with decoupled visual representations, InternVL-U integrates a state-of-the-art Multimodal Large Language Model (MLLM) with a specialized MMDiT-based visual generation head. To further bridge the gap between aesthetic generation and high-level intelligence, we construct a comprehensive data synthesis pipeline targeting high-semantic-density tasks, such as text rendering and scientific reasoning, under a reasoning-centric paradigm that leverages Chain-of-Thought (CoT) to better align abstract user intent with fine-grained visual generation details. Extensive experiments demonstrate that InternVL-U achieves a superior performance - efficiency balance. Despite using only 4B parameters, it consistently outperforms unified baseline models with over 3x larger scales such as BAGEL (14B) on various generation and editing tasks, while retaining strong multimodal understanding and reasoning capabilities.

Addressing the challenges of fragmented task definitions and the heterogeneity of unstructured data in multimodal parsing, this paper proposes the Omni Parsing framework. This framework establishes a Unified Taxonomy covering documents, images, and audio-visual streams, introducing a progressive parsing paradigm that bridges perception and cognition. Specifically, the framework integrates three hierarchical levels: 1) Holistic Detection, which achieves precise spatial-temporal grounding of objects or events to establish a geometric baseline for perception; 2) Fine-grained Recognition, which performs symbolization (e.g., OCR/ASR) and attribute extraction on localized objects to complete structured entity parsing; and 3) Multi-level Interpreting, which constructs a reasoning chain from local semantics to global logic. A pivotal advantage of this framework is its evidence anchoring mechanism, which enforces a strict alignment between high-level semantic descriptions and low-level facts. This enables ``evidence-based'' logical induction, transforming unstructured signals into standardized knowledge that is locatable, enumerable, and traceable. Building on this foundation, we constructed a standardized dataset and released the Logics-Parsing-Omni model, which successfully converts complex audio-visual signals into machine-readable structured knowledge. Experiments demonstrate that fine-grained perception and high-level cognition are synergistic, effectively enhancing model reliability. Furthermore, to quantitatively evaluate these capabilities, we introduce OmniParsingBench. Code, models and the benchmark are released at https://github.com/alibaba/Logics-Parsing/tree/master/Logics-Parsing-Omni.

Multimodal large language models (MLLMs) can process text presented as images, yet they often perform worse than when the same content is provided as textual tokens. We systematically diagnose this "modality gap" by evaluating seven MLLMs across seven benchmarks in five input modes, spanning both synthetically rendered text and realistic document images from arXiv PDFs to Wikipedia pages. We find that the gap is highly sensitive to rendering choices such as font and resolution, and that natural document images often exhibit much smaller gaps, suggesting the performance difference partly reflects evaluation artifacts rather than fundamental limitations. Through a grounded-theory error analysis of over 4,000 examples, we identify the primary cause: image input alone suppresses reasoning effort, with models producing 5--19x shorter outputs that skip step-by-step computation or reasoning. The reluctance to reason, not a failure of perception or knowledge retrieval, drives the performance gap, particularly on tasks requiring multi-step reasoning. We show that a simple, lightweight on-policy self-distillation method by fine-tuning models on their own text-mode reasoning traces paired with image inputs closes this gap, raising image-mode accuracy to match or exceed text-mode performance with over 50% improvement, and the gains transfer to unseen benchmarks without catastrophic forgetting. Overall, our results and analyses provide a systematic understanding of the modality gap and suggest a practical path toward improving visual text understanding in multimodal language models.

Multimodal large language models have recently shown promising progress in visual mathematical reasoning. However, their performance is often limited by a critical yet underexplored bottleneck: inaccurate visual perception. Through systematic analysis, we find that the most failures originate from incorrect or incomplete visual evidence extraction rather than deficiencies in reasoning capability. Moreover, models tend to remain overly confident in their initial perceptions, making standard strategies such as prompt engineering, multi-round self-reflection, or posterior guidance insufficient to reliably correct errors. To address this limitation, we propose M3-ACE, a multi-agentic context engineering framework designed to rectify visual perception in multimodal math reasoning. Instead of directly aggregating final answers, our approach decouples perception and reasoning by dynamically maintaining a shared context centered on visual evidence lists. Multiple agents collaboratively contribute complementary observations, enabling the system to expose inconsistencies and recover missing perceptual information. To support stable multi-turn collaboration, we further introduce two lightweight tools: a Summary Tool that organizes evidence from different agents into consistent, complementary, and conflicting components, and a Refine Tool that filters unreliable samples and guides iterative correction. Extensive experiments demonstrate that M3-ACE substantially improves visual mathematical reasoning performance across multiple benchmarks. Our method establishes new state-of-the-art results 89.1 on the MathVision benchmark and achieves consistent improvements on other related datasets, including MathVista and MathVerse. These results highlight the importance of perception-centric multi-agent collaboration for advancing multimodal reasoning systems.

Multimodal Mathematical Reasoning (MMR) has recently attracted increasing attention for its capability to solve mathematical problems involving both textual and visual modalities. However, current models still face significant challenges in real-world visual math tasks, often misinterpreting diagrams, failing to align mathematical symbols with visual evidence, or producing inconsistent reasoning steps. Moreover, existing evaluations mainly focus on checking final answers rather than verifying the correctness or executability of each intermediate step. A growing body of recent research addresses these issues by integrating structured perception, explicit alignment, and verifiable reasoning within unified frameworks. To establish a clear roadmap for understanding and comparing different MMR approaches, we systematically review them around four fundamental questions: (1) What to extract from multimodal inputs, (2) How to represent and align textual and visual information, (3) How to perform the reasoning, and (4) How to evaluate the correctness of the overall reasoning process. Finally, we discuss open challenges and share our thoughts on future research directions.

Multimodal large language models (MLLMs) achieve strong performance on benchmarks that evaluate text, image, or video understanding separately. However, these settings do not assess a critical real-world requirement, which involves retrieving relevant evidence from large, heterogeneous multimodal corpora prior to reasoning. Most existing benchmarks restrict retrieval to small, single-modality candidate sets, substantially simplifying the search space and overstating end-to-end reliability. To address this gap, we introduce MultiHaystack, the first benchmark designed to evaluate both retrieval and reasoning under large-scale, cross-modal conditions. MultiHaystack comprises over 46,000 multimodal retrieval candidates across documents, images, and videos, along with 747 open yet verifiable questions. Each question is grounded in a unique validated evidence item within the retrieval pool, requiring evidence localization across modalities and fine-grained reasoning. In our study, we find that models perform competitively when provided with the corresponding evidence, but their performance drops sharply when required to retrieve that evidence from the full corpus. Additionally, even the strongest retriever, E5-V, achieves only 40.8% Recall@1, while state-of-the-art MLLMs such as GPT-5 experience a significant drop in reasoning accuracy from 80.86% when provided with the corresponding evidence to 51.4% under top-5 retrieval. These results indicate that multimodal retrieval over heterogeneous pools remains a primary bottleneck for MLLMs, positioning MultiHaystack as a valuable testbed that highlights underlying limitations obscured by small-scale evaluations and promotes retrieval-centric advances in multimodal systems.

Establishing common ground, a shared set of beliefs and mutually recognized facts, is fundamental to collaboration, yet remains a challenge for current AI systems, especially in multimodal, multiparty settings, where the collaborators bring different information to the table. We introduce the Distributed Partial Information Puzzle (DPIP), a collaborative construction task that elicits rich multimodal communication under epistemic asymmetry. We present a multimodal dataset of these interactions, annotated and temporally aligned across speech, gesture, and action modalities to support reasoning over propositional content and belief dynamics. We then evaluate two paradigms for modeling common ground (CG): (1) state-of-the-art large language models (LLMs), prompted to infer shared beliefs from multimodal updates, and (2) an axiomatic pipeline grounded in Dynamic Epistemic Logic (DEL) that incrementally performs the same task. Results on the annotated DPIP data indicate that it poses a challenge to modern LLMs' abilities to track both task progression and belief state.