VLM Reasoning

Reasoning-based text-to-image (T2I) generation requires models to interpret complex prompts accurately. Existing reasoning frameworks can be broadly categorized into two types: (1) Text-Only Reasoning, which is computationally efficient but lacks access to visual context, often resulting in the omission of critical spatial and visual elements; and (2) Text-Image Interleaved Reasoning, which leverages a T2I generator to provide visual references during the reasoning process. While this approach enhances visual grounding, it incurs substantial computational costs and constrains the reasoning capacity of MLLMs to the representational limitations of the generator. To this end, we propose StruVis, a novel framework that enhances T2I generation through Thinking with Structured Vision. Instead of relying on intermediate image generation, StruVis employs text-based structured visual representations as intermediate reasoning states, thereby enabling the MLLM to effectively "perceive" visual structure within a purely text-based reasoning process. Powered by this, the reasoning potential for T2I generation of the MLLM is unlocked through structured-vision-guided reasoning. Additionally, as a generator-agnostic reasoning framework, our proposed StruVis can be seamlessly integrated with diverse T2I generators and efficiently enhance their performance in reasoning-based T2I generation. Extensive experiments demonstrate that StruVis achieves significant performance improvements on reasoning-based T2I benchmarks, e.g., a 4.61% gain on T2I-ReasonBench and a 4% gain on WISE.

Multi-view spatial reasoning remains difficult for current vision-language models. Even when multiple viewpoints are available, models often underutilize cross-view relations and instead rely on single-image shortcuts, leading to fragile performance on viewpoint transformation and occlusion-sensitive cases. We present ViewFusion, a two-stage framework that explicitly separates cross-view spatial pre-alignment from question answering. In the first stage, the model performs deliberate spatial pre-thinking to infer viewpoint relations and spatial transformations across views, forming an intermediate workspace that goes beyond a simple re-description. In the second stage, the model conducts question-driven reasoning conditioned on this workspace to produce the final prediction. We train ViewFusion with synthetic reasoning supervision followed by reinforcement learning using GRPO, which improves answer correctness while stabilizing the intended two-stage generation behavior. On MMSI-Bench, ViewFusion improves accuracy by 5.3% over Qwen3-VL-4B-Instruct, with the largest gains on examples that require genuine cross-view alignment.

Vision-Language Models (VLMs) have achieved remarkable progress on a wide range of challenging multimodal understanding and reasoning tasks. However, existing reasoning paradigms, such as the classical Chain-of-Thought (CoT), rely solely on textual information and often underutilize important visual cues. While prior work has incorporated pixel-level visual cues, these representations require precise spatial localization, introducing additional learning complexity. To address this, we propose PatchCue, a novel patch-based visual cue paradigm designed to significantly enhance the visual reasoning capabilities of VLMs. By partitioning images into patches and representing cues at the patch level, PatchCue aligns better with human perceptual habits and leverages the patch-tokenized input of modern VLMs. We train VLMs using a two-stage approach: cold-start supervised fine-tuning to output patch-level cues, followed by reinforcement learning with a process-supervised cue reward that guides intermediate visual reasoning steps. Extensive experiments on multiple VLMs and diverse benchmarks, including general visual question answering, complex reasoning, and document understanding, demonstrate that PatchCue consistently improves overall model performance. Our results show that patch-level cues outperform both pixel-level bounding boxes and point-based cues, providing a more effective and cognitively aligned visual reasoning paradigm.

Knowledge-Based Visual Question Answering (KB-VQA) requires models to answer questions about an image by integrating external knowledge, posing significant challenges due to noisy retrieval and the structured, encyclopedic nature of the knowledge base. These characteristics create a distributional gap from pretrained multimodal large language models (MLLMs), making effective reasoning and domain adaptation difficult in the post-training stage. In this work, we propose Wiki-R1, a data-generation-based curriculum reinforcement learning framework that systematically incentivizes reasoning in MLLMs for KB-VQA. Wiki-R1 constructs a sequence of training distributions aligned with the model's evolving capability, bridging the gap from pretraining to the KB-VQA target distribution. We introduce controllable curriculum data generation, which manipulates the retriever to produce samples at desired difficulty levels, and a curriculum sampling strategy that selects informative samples likely to yield non-zero advantages during RL updates. Sample difficulty is estimated using observed rewards and propagated to unobserved samples to guide learning. Experiments on two KB-VQA benchmarks, Encyclopedic VQA and InfoSeek, demonstrate that Wiki-R1 achieves new state-of-the-art results, improving accuracy from 35.5% to 37.1% on Encyclopedic VQA and from 40.1% to 44.1% on InfoSeek. The project page is available at https://artanic30.github.io/project_pages/WikiR1/.

Vision--language models (VLMs) process images as visual tokens, yet their intermediate reasoning is often carried out in text, which can be suboptimal for visually grounded radiology tasks. Radiologists instead diagnose via sequential visual search; eye-tracking captures this process as time-ordered gaze trajectories that reveal how evidence is acquired over time. We use eye-gaze as supervision to guide VLM reasoning by introducing a small set of dedicated gaze tokens. These tokens are trained to predict gaze-selected image patch indices in temporal order, encouraging the model to follow human-like evidence acquisition and integration. Experiments on MIMIC-EYE and multiple external zero-shot benchmarks show consistent gains over baselines, achieving state-of-the-art in-domain performance and improved out-of-domain robustness. These results highlight temporally ordered gaze as an effective supervision signal for learning visually grounded medical reasoning.

Multi-image reasoning remains a significant challenge for vision-language models (VLMs). We investigate a previously overlooked phenomenon: during chain-of-thought (CoT) generation, the text-to-image (T2I) attention of reasoning VLMs exhibits diffuse "pulses": sporadic and unfocused attention patterns that fail to concentrate on task-relevant images. We further reveal a systematic positional bias in attention allocation across images. Motivated by these observations, we propose PulseFocus, a training-free, inference-time method that structures CoT reasoning into interleaved plan/focus blocks with soft attention gating. By forcing the model to explicitly plan which image to examine and then gating decode-time attention to the referenced image, PulseFocus sharpens attention focus and yields consistent improvements on multi-image benchmarks like BLINK benchmark (+3.7%) and MuirBench (+1.07%).

Reasoning segmentation increasingly employs reinforcement learning to generate explanatory reasoning chains that guide Multimodal Large Language Models. While these geometric rewards are primarily confined to guiding the final localization, they are incapable of discriminating whether the reasoning process remains anchored on the referred region or strays into irrelevant context. Lacking this discriminative guidance, the model's reasoning often devolves into unfocused and verbose chains that ultimately fail to disambiguate and perceive the target in complex scenes. This suggests a need to complement the RL objective with Discriminative Perception, an ability to actively distinguish a target from its context. To realize this, we propose DPAD to compel the model to generate a descriptive caption of the referred object, which is then used to explicitly discriminate by contrasting the caption's semantic relevance to the referred object against the wider context. By optimizing for this discriminative capability, the model is forced to focus on the unique attributes of the target, leading to a more converged and efficient reasoning chain. The descriptive caption also serves as an interpretability rationale that aligns with the segmentation. Experiments on the benchmarks confirm the validity of our approach, delivering substantial performance gains, with the cIoU on ReasonSeg increasing by 3.09% and the reasoning chain length decreasing by approximately 42%. Code is available at https://github.com/mrazhou/DPAD

We present Phi-4-reasoning-vision-15B, a compact open-weight multimodal reasoning model, and share the motivations, design choices, experiments, and learnings that informed its development. Our goal is to contribute practical insight to the research community on building smaller, efficient multimodal reasoning models and to share the result of these learnings as an open-weight model that is good at common vision and language tasks and excels at scientific and mathematical reasoning and understanding user interfaces. Our contributions include demonstrating that careful architecture choices and rigorous data curation enable smaller, open-weight multimodal models to achieve competitive performance with significantly less training and inference-time compute and tokens. The most substantial improvements come from systematic filtering, error correction, and synthetic augmentation -- reinforcing that data quality remains the primary lever for model performance. Systematic ablations show that high-resolution, dynamic-resolution encoders yield consistent improvements, as accurate perception is a prerequisite for high-quality reasoning. Finally, a hybrid mix of reasoning and non-reasoning data with explicit mode tokens allows a single model to deliver fast direct answers for simpler tasks and chain-of-thought reasoning for complex problems.

Robots operating in shared human environments must not only navigate, interact, and detect their surroundings, they must also interpret and respond to dynamic, and often unpredictable, human behaviours. Although recent advances have shown promise in enhancing robotic perception and instruction-following using Vision-Language Models (VLMs), they remain limited in addressing the complexities of multimodal human-robot interactions (HRI). Motivated by this challenge, we introduce a lightweight language-to-vision feedback module that closes the loop between an LLM and the vision encoder in VLMs. The module projects image-token hidden states through a gated Multi-Layer Perceptron (MLP) back into the encoder input, prompting a second pass that reinterprets the scene under text context. We evaluate this approach on three robotics-centred tasks: navigation in a simulated environment (Habitat), sequential scene description (Mementos-Robotics), and human-intention recognition (our HRI dataset). Results show that our method improves Qwen 2.5 (7B) by $3.3\%$ (less distance), $+0.057$ description score, and $+2.93\%$ accuracy, with less than $3\%$ extra parameters; Gemma 3 (4B) and LLaVA OV 1.5 (4B) show mixed navigation results but gains $+0.111,+0.055$ and $+10.81\%,+4.79\%$ on the latter two tasks. Code is available at https://github.com/alessioGalatolo/VLM-Reasoning-for-Robotics

Humans can robustly localize visual evidence and provide grounded answers even in noisy environments by identifying critical cues and then relating them to the full context in a bottom-up manner. Inspired by this, we propose DeepScan, a training-free framework that combines Hierarchical Scanning, Refocusing, and Evidence-Enhanced Reasoning for visually grounded reasoning in Large Vision-Language Models (LVLMs). Unlike existing methods that pursue one-shot localization of complete evidence, Hierarchical Scanning performs local cue exploration and multi-scale evidence extraction to recover evidence in a bottom-up manner, effectively mitigating the impacts of distractive context. Refocusing then optimizes the localized evidence view through collaboration of LVLMs and visual experts. Finally, Evidence-Enhanced Reasoning aggregates multi-granular views via a hybrid evidence memory and yields accurate and interpretable answers. Experimental results demonstrate that DeepScan significantly boosts LVLMs in diverse visual tasks, especially in fine-grained visual understanding. It achieves 90.6% overall accuracy on V* when integrated with Qwen2.5-VL-7B. Moreover, DeepScan provides consistent improvements for LVLMs across various architectures and model scales without additional adaptation cost.

The cold-start initialization stage plays a pivotal role in training Multimodal Large Reasoning Models (MLRMs), yet its mechanisms remain insufficiently understood. To analyze this stage, we introduce the Visual Attention Score (VAS), an attention-based metric that quantifies how much a model attends to visual tokens. We find that reasoning performance is strongly correlated with VAS (r=0.9616): models with higher VAS achieve substantially stronger multimodal reasoning. Surprisingly, multimodal cold-start fails to elevate VAS, resulting in attention distributions close to the base model, whereas text-only cold-start leads to a clear increase. We term this counter-intuitive phenomenon Lazy Attention Localization. To validate its causal role, we design training-free interventions that directly modulate attention allocation during inference, performance gains of 1$-$2% without any retraining. Building on these insights, we further propose Attention-Guided Visual Anchoring and Reflection (AVAR), a comprehensive cold-start framework that integrates visual-anchored data synthesis, attention-guided objectives, and visual-anchored reward shaping. Applied to Qwen2.5-VL-7B, AVAR achieves an average gain of 7.0% across 7 multimodal reasoning benchmarks. Ablation studies further confirm that each component of AVAR contributes step-wise to the overall gains. The code, data, and models are available at https://github.com/lrlbbzl/Qwen-AVAR.

Geo-temporal understanding, the ability to infer location, time, and contextual properties from visual input alone, underpins applications such as disaster management, traffic planning, embodied navigation, world modeling, and geography education. Although recent vision-language models (VLMs) have advanced image geo-localization using cues like landmarks and road signs, their ability to reason about temporal signals and physically grounded spatial cues remains limited. To address this gap, we introduce TimeSpot, a benchmark for evaluating real-world geo-temporal reasoning in VLMs. TimeSpot comprises 1,455 ground-level images from 80 countries and requires structured prediction of temporal attributes (season, month, time of day, daylight phase) and geographic attributes (continent, country, climate zone, environment type, latitude-longitude) directly from visual evidence. It also includes spatial-temporal reasoning tasks that test physical plausibility under real-world uncertainty. Evaluations of state-of-the-art open- and closed-source VLMs show low performance, particularly for temporal inference. While supervised fine-tuning yields improvements, results remain insufficient, highlighting the need for new methods to achieve robust, physically grounded geo-temporal understanding TimeSpot is available at: https://TimeSpot-GT.github.io.

We present VB, a benchmark that tests whether vision-language models can determine what is and is not visible in a photograph, and abstain when a human viewer cannot reliably answer. Each item pairs a single photo with a short yes/no visibility claim; the model must output VISIBLY_TRUE, VISIBLY_FALSE, or ABSTAIN, together with a confidence score. Items are organized into 100 families using a 2x2 design that crosses a minimal image edit with a minimal text edit, yielding 300 headline evaluation cells. Unlike prior unanswerable-VQA benchmarks, VB tests not only whether a question is unanswerable but why (via reason codes tied to specific visibility factors), and uses controlled minimal edits to verify that model judgments change when and only when the underlying evidence changes. We score models on confidence-aware accuracy with abstention (CAA), minimal-edit flip rate (MEFR), confidence-ranked selective prediction (SelRank), and second-order perspective reasoning (ToMAcc); all headline numbers are computed on the strict XOR subset (three cells per family, 300 scored items per model). We evaluate nine models spanning flagship and prior-generation closed-source systems, and open-source models from 8B to 12B parameters. GPT-4o and Gemini 3.1 Pro effectively tie for the best composite score (0.728 and 0.727), followed by Gemini 2.5 Pro (0.678). The best open-source model, Gemma 3 12B (0.505), surpasses one prior-generation closed-source system. Text-flip robustness exceeds image-flip robustness for six of nine models, and confidence calibration varies substantially: GPT-4o and Gemini 2.5 Pro achieve similar accuracy yet differ sharply in selective prediction quality.

Vision-Language Navigation (VLN) presents a unique challenge for Large Vision-Language Models (VLMs) due to their inherent architectural mismatch: VLMs are primarily pretrained on static, disembodied vision-language tasks, which fundamentally clash with the dynamic, embodied, and spatially-structured nature of navigation. Existing large-model-based methods often resort to converting rich visual and spatial information into text, forcing models to implicitly infer complex visual-topological relationships or limiting their global action capabilities. To bridge this gap, we propose TagaVLM (Topology-Aware Global Action reasoning), an end-to-end framework that explicitly injects topological structures into the VLM backbone. To introduce topological edge information, Spatial Topology Aware Residual Attention (STAR-Att) directly integrates it into the VLM's self-attention mechanism, enabling intrinsic spatial reasoning while preserving pretrained knowledge. To enhance topological node information, an Interleaved Navigation Prompt strengthens node-level visual-text alignment. Finally, with the embedded topological graph, the model is capable of global action reasoning, allowing for robust path correction. On the R2R benchmark, TagaVLM achieves state-of-the-art performance among large-model-based methods, with a Success Rate (SR) of 51.09% and SPL of 47.18 in unseen environments, outperforming prior work by 3.39% in SR and 9.08 in SPL. This demonstrates that, for embodied spatial reasoning, targeted enhancements on smaller open-source VLMs can be more effective than brute-force model scaling. The code can be found on our project page: https://apex-bjut.github.io/Taga-VLM

Large Vision Language Models (LVLMs) exhibit strong Chain-of-Thought (CoT) capabilities, yet most existing paradigms assume full-video availability before inference, a batch-style process misaligned with real-world video streams where information arrives sequentially. Motivated by the streaming nature of video data, we investigate two streaming reasoning paradigms for LVLMs. The first, an interleaved paradigm, alternates between receiving frames and producing partial reasoning but remains constrained by strictly ordered cache updates. To better match streaming inputs, we propose Think-as-You-See (TaYS), a unified framework enabling true concurrent reasoning. TaYS integrates parallelized CoT generation, stream-constrained training, and stream-parallel inference. It further employs temporally aligned reasoning units, streaming attention masks and positional encodings, and a dual KV-cache that decouples visual encoding from textual reasoning. We evaluate all paradigms on the Qwen2.5-VL family across representative video CoT tasks, including event dynamics analysis, causal reasoning, and thematic understanding. Experiments show that TaYS consistently outperforms both batch and interleaved baselines, improving reasoning performance while substantially reducing time-to-first-token (TTFT) and overall reasoning delay. These results demonstrate the effectiveness of data-aligned streaming reasoning in enabling efficient and responsive video understanding for LVLMs. We release our code at https://github.com/EIT-NLP/StreamingLLM/tree/main/TaYS

Large vision-language models (LVLMs) demonstrate strong performance on diagram understanding benchmarks, yet they still struggle with understanding relationships between elements, particularly those represented by nodes and directed edges (e.g., arrows and lines). To investigate the underlying causes of this limitation, we probe the internal representation of LVLMs using a carefully constructed synthetic diagram dataset based on directed graphs. Our probing experiments reveal that edge information is not linearly separable in the vision encoder and becomes linearly encoded only in the text tokens in the language model. In contrast, node information and global structural features are already linearly encoded in individual hidden states of the vision encoder. These findings suggest that the stage at which linearly separable representations are formed varies depending on the type of visual information. In particular, the delayed emergence of edge representations may help explain why LVLMs struggle with relational understanding, such as interpreting edge directions, which require more abstract, compositionally integrated processes.

Despite the success of Large Vision--Language Models (LVLMs), most existing architectures suffer from a representation bottleneck: they rely on static, instruction-agnostic vision encoders whose visual representations are utilized in an invariant manner across different textual tasks. This rigidity hinders fine-grained reasoning where task-specific visual cues are critical. To address this issue, we propose iGVLM, a general framework for instruction-guided visual modulation. iGVLM introduces a decoupled dual-branch architecture: a frozen representation branch that preserves task-agnostic visual representations learned during pre-training, and a dynamic conditioning branch that performs affine feature modulation via Adaptive Layer Normalization (AdaLN). This design enables a smooth transition from general-purpose perception to instruction-aware reasoning while maintaining the structural integrity and stability of pre-trained visual priors. Beyond standard benchmarks, we introduce MM4, a controlled diagnostic probe for quantifying logical consistency under multi-query, multi-instruction settings. Extensive results show that iGVLM consistently enhances instruction sensitivity across diverse language backbones, offering a plug-and-play paradigm for bridging passive perception and active reasoning.

Vision-language models remain susceptible to multimodal jailbreaks and over-refusal because safety hinges on both visual evidence and user intent, while many alignment pipelines supervise only the final response. To address this, we present SaFeR-ToolKit, which formalizes safety decision-making as a checkable protocol. Concretely, a planner specifies a persona, a Perception $\to$ Reasoning $\to$ Decision tool set, and a constrained transition graph, while a responder outputs a typed key-value tool trace before the final answer. To make the protocol reliably followed in practice, we train a single policy with a three-stage curriculum (SFT $\to$ DPO $\to$ GRPO), where GRPO directly supervises tool usage beyond answer-level feedback. Our contributions are two-fold: I. Dataset. The first tool-based safety reasoning dataset, comprising 31,654 examples (SFT 6k, DPO 18.6k, GRPO 6k) plus 1k held-out evaluation. II. Experiments. On Qwen2.5-VL, SaFeR-ToolKit significantly improves Safety/Helpfulness/Reasoning Rigor on 3B (29.39/45.04/4.98 $\to$ 84.40/71.13/78.87) and 7B (53.21/52.92/19.26 $\to$ 86.34/80.79/85.34), while preserving general capabilities (3B: 58.67 $\to$ 59.21; 7B: 66.39 $\to$ 66.81). Codes are available at https://github.com/Duebassx/SaFeR_ToolKit.

Reasoning has emerged as a key capability of large language models. In linguistic tasks, this capability can be enhanced by self-improving techniques that refine reasoning paths for subsequent finetuning. However, extending these language-based self-improving approaches to vision language models (VLMs) presents a unique challenge:~visual hallucinations in reasoning paths cannot be effectively verified or rectified. Our solution starts with a key observation about visual contrast: when presented with a contrastive VQA pair, i.e., two visually similar images with synonymous questions, VLMs identify relevant visual cues more precisely. Motivated by this observation, we propose Visual Contrastive Self-Taught Reasoner (VC-STaR), a novel self-improving framework that leverages visual contrast to mitigate hallucinations in model-generated rationales. We collect a diverse suite of VQA datasets, curate contrastive pairs according to multi-modal similarity, and generate rationales using VC-STaR. Consequently, we obtain a new visual reasoning dataset, VisCoR-55K, which is then used to boost the reasoning capability of various VLMs through supervised finetuning. Extensive experiments show that VC-STaR not only outperforms existing self-improving approaches but also surpasses models finetuned on the SoTA visual reasoning datasets, demonstrating that the inherent contrastive ability of VLMs can bootstrap their own visual reasoning. Project at: https://github.com/zhiyupan42/VC-STaR.

Think-Answer reasoners such as DeepSeek-R1 have made notable progress by leveraging interpretable internal reasoning. However, despite the frequent presence of self-reflective cues like "Oops!", they remain vulnerable to output errors during single-pass inference. To address this limitation, we propose an efficient Recursive Think-Answer Process (R-TAP) that enables models to engage in iterative reasoning cycles and generate more accurate answers, going beyond conventional single-pass approaches. Central to this approach is a confidence generator that evaluates the certainty of model responses and guides subsequent improvements. By incorporating two complementary rewards-Recursively Confidence Increase Reward and Final Answer Confidence Reward-we show that R-TAP-enhanced models consistently outperform conventional single-pass methods for both large language models (LLMs) and vision-language models (VLMs). Moreover, by analyzing the frequency of "Oops"-like expressions in model responses, we find that R-TAP-applied models exhibit significantly fewer self-reflective patterns, resulting in more stable and faster inference-time reasoning. We hope R-TAP pave the way evolving into efficient and elaborated methods to refine the reasoning processes of future AI.

Large vision-language models (VLMs) often benefit from chain-of-thought (CoT) prompting in general domains, yet its efficacy in medical vision-language tasks remains underexplored. We report a counter-intuitive trend: on medical visual question answering, CoT frequently underperforms direct answering (DirA) across general-purpose and medical-specific models. We attribute this to a medical perception bottleneck: subtle, domain-specific cues can weaken visual grounding, and CoT may compound early perceptual uncertainty rather than correct it. To probe this hypothesis, we introduce two training-free, inference-time grounding interventions: (i) perception anchoring via region-of-interest cues and (ii) description grounding via high-quality textual guidance. Across multiple benchmarks and model families, these interventions improve accuracy, mitigate CoT degradation, and in several settings reverse the CoT--DirA inversion. Our findings suggest that reliable clinical VLMs require robust visual grounding and cross-modal alignment, beyond extending text-driven reasoning chains. Code is available here.

Recent advances in training-free visual prompting, such as Set-of-Mark, have emerged as a promising direction for enhancing the grounding capabilities of multimodal language models (MLMs). These techniques operate by partitioning the input image into object regions and annotating them with marks, predominantly boxes with numeric identifiers, before feeding the augmented image to the MLM. However, these approaches treat marked objects as isolated entities, failing to capture the relationships between them. On these premises, we propose Graph-of-Mark (GoM), the first pixel-level visual prompting technique that overlays scene graphs onto the input image for spatial reasoning tasks. We evaluate GoM across 3 open-source MLMs and 4 different datasets, conducting extensive ablations on drawn components and investigating the impact of auxiliary graph descriptions in the text prompt. Our results demonstrate that GoM consistently improves the zero-shot capability of MLMs in interpreting object positions and relative directions, improving base accuracy in visual question answering and localization up to 11 percentage points.

Hateful memes often require compositional multimodal reasoning: the image and text may appear benign in isolation, yet their interaction conveys harmful intent. Although thinking-based multimodal large language models (MLLMs) have recently advanced vision-language understanding, their capabilities remain underexplored for hateful meme analysis. We propose a reinforcement learning based post-training framework that improves reasoning in thinking-based MLLMs through task-specific rewards and a novel Group Relative Policy Optimization (GRPO) objective. Specifically, we (i) conduct a systematic empirical study of off-the-shelf MLLMs for hateful meme understanding, (ii) extend an existing hateful meme dataset by generating weakly or pseudo-supervised chain-of-thought rationales via distillation, and (iii) introduce a GRPO-based objective that jointly optimizes meme classification and explanation quality to encourage fine-grained, step-by-step reasoning. Experiments on the Hateful Memes benchmark show that our approach achieves state-of-the-art performance, improving accuracy and F1 by approximately 1 percent and explanation quality by approximately 3 percent. We will publicly release our code, dataset extensions, and evaluation resources to support reproducibility.

Humans are born with vision-based 4D spatial-temporal intelligence, which enables us to perceive and reason about the evolution of 3D space over time from purely visual inputs. Despite its importance, this capability remains a significant bottleneck for current multimodal large language models (MLLMs). To tackle this challenge, we introduce MLLM-4D, a comprehensive framework designed to bridge the gaps in training data curation and model post-training for spatiotemporal understanding and reasoning. On the data front, we develop a cost-efficient data curation pipeline that repurposes existing stereo video datasets into high-quality 4D spatiotemporal instructional data. This results in the MLLM4D-2M and MLLM4D-R1-30k datasets for Supervised Fine-Tuning (SFT) and Reinforcement Fine-Tuning (RFT), alongside MLLM4D-Bench for comprehensive evaluation. Regarding model training, our post-training strategy establishes a foundational 4D understanding via SFT and further catalyzes 4D reasoning capabilities by employing Group Relative Policy Optimization (GRPO) with specialized Spatiotemporal Chain of Thought (ST-CoT) prompting and Spatiotemporal reward functions (ST-reward) without involving the modification of architecture. Extensive experiments demonstrate that MLLM-4D achieves state-of-the-art spatial-temporal understanding and reasoning capabilities from purely 2D RGB inputs. Project page: https://github.com/GVCLab/MLLM-4D.

We present ReMoT, a unified training paradigm to systematically address the fundamental shortcomings of VLMs in spatio-temporal consistency -- a critical failure point in navigation, robotics, and autonomous driving. ReMoT integrates two core components: (1) A rule-based automatic framework that generates ReMoT-16K, a large-scale (16.5K triplets) motion-contrast dataset derived from video meta-annotations, surpassing costly manual or model-based generation. (2) Group Relative Policy Optimization, which we empirically validate yields optimal performance and data efficiency for learning this contrastive reasoning, far exceeding standard Supervised Fine-Tuning. We also construct the first benchmark for fine-grained motion contrast triplets to measure a VLM's discrimination of subtle motion attributes (e.g., opposing directions). The resulting model achieves state-of-the-art performance on our new benchmark and multiple standard VLM benchmarks, culminating in a remarkable 25.1% performance leap on spatio-temporal reasoning tasks.

Vision-language models (VLMs) show promise in drafting radiology reports, yet they frequently suffer from logical inconsistencies, generating diagnostic impressions unsupported by their own perceptual findings or missing logically entailed conclusions. Standard lexical metrics heavily penalize clinical paraphrasing and fail to capture these deductive failures in reference-free settings. Toward guarantees for clinical reasoning, we introduce a neurosymbolic verification framework that deterministically audits the internal consistency of VLM-generated reports. Our pipeline autoformalizes free-text radiographic findings into structured propositional evidence, utilizing an SMT solver (Z3) and a clinical knowledge base to verify whether each diagnostic claim is mathematically entailed, hallucinated, or omitted. Evaluating seven VLMs across five chest X-ray benchmarks, our verifier exposes distinct reasoning failure modes, such as conservative observation and stochastic hallucination, that remain invisible to traditional metrics. On labeled datasets, enforcing solver-backed entailment acts as a rigorous post-hoc guarantee, systematically eliminating unsupported hallucinations to significantly increase diagnostic soundness and precision in generative clinical assistants.

Knowledge-based visual question answering (KB-VQA) demonstrates significant potential for handling knowledge-intensive tasks. However, conflicts arise between static parametric knowledge in vision language models (VLMs) and dynamically retrieved information due to the static model knowledge from pre-training. The outputs either ignore retrieved contexts or exhibit inconsistent integration with parametric knowledge, posing substantial challenges for KB-VQA. Current knowledge conflict mitigation methods primarily adapted from language-based approaches, focusing on context-level conflicts through engineered prompting strategies or context-aware decoding mechanisms. However, these methods neglect the critical role of visual information in conflicts and suffer from redundant retrieved contexts, which impair accurate conflict identification and effective mitigation. To address these limitations, we propose CC-VQA: a novel training-free, conflict- and correlation-aware method for KB-VQA. Our method comprises two core components: (1) Vision-Centric Contextual Conflict Reasoning, which performs visual-semantic conflict analysis across internal and external knowledge contexts; and (2) Correlation-Guided Encoding and Decoding, featuring positional encoding compression for low-correlation statements and adaptive decoding using correlation-weighted conflict scoring. Extensive evaluations on E-VQA, InfoSeek, and OK-VQA benchmarks demonstrate that CC-VQA achieves state-of-the-art performance, yielding absolute accuracy improvements of 3.3% to 6.4% compared to existing methods. Code is available at https://github.com/cqu-student/CC-VQA.

Recent advancements in Multimodal Large Language Models (MLLMs) pursue omni-perception capabilities, yet integrating robust sensory grounding with complex reasoning remains a challenge, particularly for underrepresented regions. In this report, we introduce the research preview of MERaLiON2-Omni (Alpha), a 10B-parameter multilingual omni-perception tailored for Southeast Asia (SEA). We present a progressive training pipeline that explicitly decouples and then integrates "System 1" (Perception) and "System 2" (Reasoning) capabilities. First, we establish a robust Perception Backbone by aligning region-specific audio-visual cues (e.g., Singlish code-switching, local cultural landmarks) with a multilingual LLM through orthogonal modality adaptation. Second, to inject cognitive capabilities without large-scale supervision, we propose a cost-effective Generate-Judge-Refine pipeline. By utilizing a Super-LLM to filter hallucinations and resolve conflicts via a consensus mechanism, we synthesize high-quality silver data that transfers textual Chain-of-Thought reasoning to multimodal scenarios. Comprehensive evaluation on our newly introduced SEA-Omni Benchmark Suite reveals an Efficiency-Stability Paradox: while reasoning acts as a non-linear amplifier for abstract tasks (boosting mathematical and instruction-following performance significantly), it introduces instability in low-level sensory processing. Specifically, we identify Temporal Drift in long-context audio, where extended reasoning desynchronizes the model from acoustic timestamps, and Visual Over-interpretation, where logic overrides pixel-level reality. This report details the architecture, the data-efficient training recipe, and a diagnostic analysis of the trade-offs between robust perception and structured reasoning.

Unified Multimodal Large Language Models (U-MLLMs) integrate understanding and generation within a single architecture. However, existing evaluations typically assess these capabilities separately, overlooking semantic equivalence, i.e., the ability to manifest consistent reasoning results regardless of the output modality. In this work, we investigate whether current U-MLLMs satisfy this premise. We observe that while models demonstrate robust textual reasoning, they fail to maintain semantic equivalence when required to render the same results in the image modality. To rigorously diagnose this discrepancy, we introduce VGUBench, a framework to decouple reasoning logic from generation fidelity. VGUBench comprises three diagnostic tasks: (1)Textual Generative Understanding, establishing a baseline for reasoning accuracy in textual response; (2)Visual Generative Understanding, evaluating the ability to generate visual responses that represent the correct answer; and (3)a Visual Rendering control task, which assesses the ability to directly render explicit visual descriptions into images without complex reasoning. Our evaluation reveals a significant disparity: despite strong performance in textual understanding and visual rendering, U-MLLMs exhibit a marked performance collapse when required to generate visual answers to questions. Furthermore, we find a negligible correlation between visual answering performance and basic rendering quality. These results suggest that the failure stems not from insufficient generation fidelity, but from a breakdown in cross-modal semantic alignment. We provide diagnostic insights to address this challenge in future Unified Generation and Understanding Models.

Current Large Multimodal Models (LMMs) struggle with high-resolution visual inputs during the reasoning process, as the number of image tokens increases quadratically with resolution, introducing substantial redundancy and irrelevant information. A common practice is to identify key image regions and refer to their high-resolution counterparts during reasoning, typically trained with external visual supervision. However, such visual supervision cues require costly grounding labels from human annotators. Meanwhile, it remains an open question how to enhance a model's grounding abilities to support reasoning without relying on additional annotations. In this paper, we propose High-resolution Annotation-free Reasoning Technique (HART), a closed-loop framework that enables LMMs to focus on and self-verify key regions of high-resolution visual inputs. HART incorporates a post-training paradigm in which we design Advantage Preference Group Relative Policy Optimization (AP-GRPO) to encourage accurate localization of key regions without external visual annotations. Notably, HART provides explainable reasoning pathways and enables efficient optimization of localization. Extensive experiments on MME-RealWorld-Lite, TreeBench, V* Bench, HR-Bench-4K/8K, and MMStar demonstrate that HART improves performance across a wide range of high-resolution visual tasks, consistently outperforming strong baselines.