Multimodal Model Evaluation

Follow this topic Ask about this topic

Momentum

49 papers in the last four weeks, up 44% on the four weeks before. 0.5% of all new papers.

Jul 13Week of Sep 28

Latest papers 496

Jun 24, 2026cs.CL

Same Evidence, Different Answer: Auditing Order Sensitivity in Multimodal Large Language Models

Standard benchmarks for multimodal large language models (MLLMs) score each item on one canonical ordering and miss whether order-irrelevant shuffling changes the answer, a baseline reliability property called for by emerging AI evaluation guidelines. We introduce Facet-Probe, a five-facet audit (option, evidence-chunk, document-rank, image-set, and mixed-modality ordering) of 18 frontier and open-weight MLLMs. A Bayesian item-response model separates ordering noise from per-facet bias, and a same-ordering control estimates the decoder-stochastic floor for observed flips. We find that none of the 18 MLLMs we audit are order-invariant: screened per-facet panel-mean flip rates span 24-50%. A Gemini same-ordering control at temperature 0 estimates a substantial ordering excess over a same-input decoder-noise floor in verified cells. Capability predicts but does not eliminate flips; the best model still flips on 13.4% of trials. In our Gemini mitigation tests, training-free prompt changes are modality-conditional and do not transfer from text to visual reasoning. These results suggest that prompt-level mitigation alone is unlikely to provide general order robustness, motivating future work on training-time and architectural approaches. We propose cross-ordering flip rate as a standard reporting axis for MLLMs.

Akshay Paruchuri, Sanmi Koyejo, Ehsan Adeli

Stanford University

Multimodal Robustness LLM Auditing Multimodal Model Evaluation Multimodal Understanding

Jun 24, 2026cs.CV

TriViewBench: Controlled Complexity Scaling for Multi-View Structural Reasoning in MLLMs

Multimodal Large Language Models (MLLMs) demonstrate strong performance on standard visual question answering benchmarks, yet their scalability under controlled structural complexity remains poorly understood. We introduce TriViewBench, a controlled three-view visual reasoning benchmark constructed from synthetic 3D scenes with explicitly parameterized object count and occlusion. The benchmark contains 1,923 scenes and over 14K Question-Answer (QA) pairs organized into four complexity levels and three reasoning categories: Local Decision, Object Counting, and Global Recovery. We evaluate 18 open- and closed-source MLLMs under a unified prompting protocol. All 18 models exhibit an identical capability hierarchy without exception (Local Decision > Object Counting > Global Recovery), and performance degrades monotonically with complexity: Local Decision tasks decline modestly (12.11% relative drop), while Object Counting degrades substantially (59.14%) and Global Recovery collapses severely (80.02%). Error analysis on Object Counting reveals two mechanistically independent failure modes: single-view tasks are dominated by undercounting due to occlusion blindness, whereas the multi-view task reverses to overcounting due to cross-view identity confusion. Chain-of-Thought (CoT) prompting yields near-zero overall benefit (

Δ= -0.16\%

) and its effect on Global Recovery is strongly capability-gated, suggesting that the bottleneck lies in cross-view spatial representation rather than reasoning strategy. These findings reveal fundamental scalability limitations in current MLLMs and position TriViewBench as a controlled diagnostic framework for analyzing structural reasoning failures.

Yu-Yang Chen, Lan-Zhe Guo

School of Intelligence Science and Technology, Nanjing University, China · National Key Laboratory for Novel Software Technology, Nanjing University, China

Spatial Reasoning Benchmarks Visual Reasoning Object Counting Multimodal Reasoning Visually Grounded Reasoning Multimodal Model Evaluation+2

Jun 24, 2026cs.LG

Can AI Draw Science? A Benchmark for Evaluating Scientific Figure Generation by Text-to-Image and Multimodal Models

Text-to-image and multimodal generative models are increasingly used to produce scientific figures such as mechanism diagrams, experimental-design schematics, conceptual frameworks, and graphical abstracts. Yet existing image-generation benchmarks (e.g., GenEval, T2I-CompBench, DPG-Bench) evaluate natural images and measure compositionality, object counting, or photorealism. None of them measure what makes a generated scientific figure usable: correct and legible text labels, faithful depiction of entities and their relations, coherent diagrammatic structure, and adherence to disciplinary drawing conventions. We introduce SciDraw-Bench, a benchmark of 32 structured scientific-figure generation tasks spanning eight figure types and ten disciplines, where each task pairs a natural-language prompt with a machine-checkable specification of required labels, relations, components, conventions, and negative constraints. We propose a four-dimensional evaluation protocol: Text Fidelity (OCR-based label recall and character error rate), Semantic Correctness (vision-language-model judging against the specification), Structural Quality, and Convention Adherence, together with a meta-evaluation protocol and a preliminary inter-judge reliability analysis (human-rating validation is ongoing). We evaluate a domain-specific system, SciDraw AI, against representative general-purpose text-to-image models, and outline a code-to-figure baseline as a planned extension. In a pilot over all eight figure types, the domain-specific system substantially outperforms the general-purpose baselines on every dimension and figure type, with the largest gaps on semantic correctness and convention adherence; text fidelity remains the hardest dimension for all systems.

Davie Chen

Scientific Visualization Multimodal Model Evaluation T2I Generation Evaluation

Jun 24, 2026cs.CV

Falcon: Functional Assembly and Language for Compositional Reasoning in X-ray

Conventional vision-language models are largely object-centric, focusing on detecting and describing individual entities. In safety-critical X-ray baggage screening, however, threat often emerges not from a single object but from the functional compatibility of spatially dispersed components, such as batteries, detonators, and explosive charges. We formalize this setting as \emph{compositional threat reasoning}, where risk is modeled as a relational property of grounded regions rather than an independent detection outcome. We introduce \textbf{Falcon}, a multimodal framework that abstracts segmentation-aware region features into a structured safety state capturing component presence, pairwise functional compatibility, and scene-level risk. This structured representation is injected into the language model as an explicit intermediate interface, encouraging relationally consistent and safety-aware reasoning. To evaluate this problem, we present \textbf{Falcon-X}, a benchmark that unifies dense grounding with structured supervision over component completeness and risk inference in cluttered X-ray imagery. Experiments show that while existing multimodal models adapt to appearance, they struggle with compositional safety reasoning. Falcon improves functional grounding and produces more coherent threat assessments, establishing compositional safety reasoning as a distinct evaluation paradigm for multimodal systems.

Yonathan Michael, Mohamad Alansari, Natnael Takele +2

Computer Science Department Khalifa University · Center of Cyber-physical Systems (C2PS) Khalifa University

Multimodal Grounding Multimodal Reasoning VLM Reasoning Multimodal Model Evaluation

Jun 24, 2026cs.CV

SSMNBench: Diagnosing Image-based Cross-View Human-Object Understanding via Single-View Sufficiency and Multi-View Necessity

Multimodal Large Language Models (MLLMs) have shown remarkable progress in single-image perception, yet their ability to reason about complex cross-view human-centric scenes remains largely unverified. Current multi-view benchmarks evaluate models using a fixed "bag of frames" and thus conflate a model's robustness to visual distraction with its genuine ability to fuse fragmented cross-view evidence. To address this issue, we introduce SSMNBench, a diagnostic benchmark comprising 3,300 curated QA pairs for cross-view human and human-object understanding. SSMNBench uniquely categorizes tasks into Single-View Sufficiency (SVS) and Multi-View Necessity (MVN). By systematically perturbing view availability across 17 state-of-the-art MLLMs, critical limitations are revealed: models suffer from severe "distraction degradation" when presented with redundant views (SVS), and fail to integrate fragmented geometric evidence across cameras (MVN). Our evaluations demonstrate that modern MLLMs rely on multiple single-image semantic averaging and view preference rather than genuine cross-view synthesis. By exposing these fundamental vulnerabilities, SSMNBench provides a rigorous diagnostic framework to drive the advancement of future cross-view-aware multimodal architectures. The code is available at:

\href{https://github.com/gtc-gh/SSMNBench}{\text{SSMNBench}}

Tianchen Guo, Chen Liu, Ling Chen +1

The University of Queensland · Follow Me AI Pty LTD · University of Technology Sydney +1

Visual Question Answering Multi-View Consistency Visual Reasoning Multimodal Model Evaluation Multimodal Understanding

Jun 23, 2026cs.CV

Yuvion VL: A Multimodal Foundation Model for Adversarial Content and AI Safety

General-purpose models often struggle to reliably identify and understand real-world multimodal risks, largely due to the inherent multimodal adversarial nature of content and AI safety. We present Yuvion VL, a family of multimodal large language models purpose-built for content and AI safety, with both instruction-tuned and reasoning-oriented variants. Yuvion VL addresses this gap by treating safety as an inherently adversarial and multimodal problem and designing the entire pipeline around adversarial robustness. For data construction, we develop an automated pipeline integrating adversarial-aware data synthesis with multi-stage quality control, producing large-scale, high-quality multimodal samples augmented with domain knowledge and reasoning annotations. For training, we adopt a three-stage pipeline that includes continued pretraining for risk-concept cross-modal alignment, instruct post-training for production-grade safety tasks, and reasoning post-training for enhanced interpretability and performance in complex tasks. We further introduce Confuse-then-Contrast Fine-Tuning, a contrastive framework that mines model-specific confusions and constructs multi-image contrastive groups to enforce explicit discrimination of fine-grained visual-semantic elements, enabling the model to distinguish between visually similar cases with different safety implications in adversarial safety tasks. To support rigorous evaluation, we further introduce Yuvion VL RiskEval (YVRE), a collection of benchmarks covering diverse open and internal evaluations, with a focus on content and AI safety, adversarial robustness, and real-world capability requirements. Experiments show that Yuvion VL-32B achieves industry-leading safety performance, surpassing comparably sized open-source models and best closed-source commercial models, while maintaining comparable general capabilities.

Shikai Qiu, Xiaowen Xu, Benlei Cui +55

VLM Robustness Multimodal Large Language Models LLM Safety Multimodal Model Evaluation

Jun 23, 2026cs.CL

MedBench v5: A Dynamic, Process-Oriented, and Hallucination-Aware Benchmark for Clinical Multimodal Models

Existing medical AI benchmarks lack process visibility, atomic skill evaluation, and integrated hallucination detection. We introduce MedBench v5, a redesigned benchmark for clinical multimodal models (language, vision-language, and agent systems) that moves from static QA to dynamic, process-oriented evaluation. MedBench v5 features: (1) a dual-dimensional framework combining Clinical Cognitive Responsiveness (13 sub-dimensions) and Medical Atomic Skills (4 agent environments), covering 63 tasks; (2) three switchable information-flow stressors (omission, contradiction, evidence delay) for factorized degradation analysis; (3) a dynamic process audit protocol with five reasoning nodes that produces model-specific failure fingerprints; (4) hallucination propagation monitoring across initiation, propagation, anchoring, and contradiction interaction-capturing silent hallucination. Experiments on frontier models show that strong overall task performance does not guarantee process stability: stressors mainly disrupt contradiction detection, diagnosis updating, hallucination propagation, and contradiction-based self-correction, while final evidence grounding can remain superficially stable. MedBench v5 provides a unified infrastructure for capability profiling, controllable stress testing, process auditing, and hallucination trajectory analysis in clinical AI evaluation.

Jinru Ding, Chuchu Jiang, Lu Lu +12

1Shanghai Artificial Intelligence Laboratory

AI Agent Benchmarks Medical VLMs LLM Hallucination Detection Clinical Language Model Evaluation Multimodal Model Evaluation

Jun 23, 2026cs.AI

ReMMD: Realistic Multilingual Multi-Image Agentic Verification for Multimodal Misinformation Detection

Multimodal misinformation detection is increasingly important because viral posts now combine long multilingual narratives, several images, mixed provenance, and subtle text--image framing errors. Existing benchmarks and methods remain poorly matched to this setting: they usually isolate short captions, single images, binary labels, or one manipulation source, while agentic verification remains costly under realistic evidence search. We present ReMMD, a realistic multilingual multi-image agentic verification framework for multimodal misinformation detection. ReMMD includes ReMMDBench, a real-world multimodal misinformation detection benchmark with 500 samples, 2,756 images, five monolingual languages, two cross-lingual settings, three text-length tiers, multi-image posts, five-way veracity labels, eight distortion labels, evidence provenance, and rationales. It also includes ReMMD-Agent, a persistent-memory verifier that decomposes posts into atomic points, builds a reusable evidence set, and predicts structured L1/L2/L3 outputs. Across proprietary systems, open LVLMs, MMD-Agent, and T2-Agent, ReMMD-Agent obtains the best five-way veracity performance, with 41.80% accuracy and 39.12% macro-F1 using GPT-5.2, while reducing cost by 17.5% relative to MMD-Agent and 79.9% relative to T2-Agent. The project is available at https://dang-ai.github.io/ReMMD.

Chenhao Dang, Dantong Zhu, Jun Yang +2

Shanghai Jiaotong University · Shanghai Artificial Intelligence Laboratory · Central South University +2

Multimodal Agents Multimodal Model Evaluation

Jun 22, 2026cs.CL

PIVOTSBench: Evaluating Fine-Grained Interpersonal Relationship Reasoning in Multimodal Large Language Models

Humans possess an innate ability to understand fine-grained interpersonal relationships, which is central to everyday social interactions. Although such reasoning is inherently multimodal, it remains largely unexplored by existing multimodal large language models (MLLMs). To address this gap, we introduce PIVOTS, the first benchmark built from Social-IQ 2.0 and YouTube data to evaluate MLLMs' ability to predict bidirectional interpersonal relationship dimensions grounded in established psychology research. In addition, PIVOTS includes auxiliary tasks that assess models' ability to identify and leverage the critical visual cues underlying such predictions. We evaluate both proprietary and open-source MLLMs and conduct detailed ablation studies to analyze the effects of visual modalities and explicit social role information in conversational utterances. We further examine how joint and pairwise prediction settings benefit MLLMs in scoring bidirectional PIVOTS dimensions. Project page: https://ciossayin.github.io/pivots-bench/

Shuxiang Zhang, Yiting Yin, Wenxuan Song +2

Sun Yat-sen University · University of Michigan · Tsinghua University

Multimodal Large Language Models Multimodal Reasoning Social Reasoning Multimodal Model Evaluation

Jun 22, 2026cs.CV

SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning

Vision-language models (VLMs) are increasingly deployed in consumer, medical, financial, and enterprise applications. This broad deployment expands the safety surface: risks can arise from multimodal question answering, assistant responses, and cross-modal composition, while moderation policies may vary across products, regions, and deployment stages. Most existing guardrails either rely on fixed taxonomies or target only a narrow set of interaction settings, which limits their adaptability when safety rules change at deployment time. We present \textbf{SingGuard}, a policy-adaptive multimodal guardrail model family for safety assessment in multimodal conversations. SingGuard treats the active policy as a runtime input: given natural-language rules, it checks the target content against the active policy rule by rule and predicts both the safety label and the triggered rule. To balance efficiency and interpretability, SingGuard supports fast, hybrid, and slow inference regimes along a fast-to-slow reasoning spectrum, ranging from direct safety judgments to policy-grounded deliberation. We further optimize this behavior with fast--slow decoupled reinforcement learning. We also introduce \textbf{SingGuard-Bench}, a multimodal guardrail benchmark with 56{,}340 examples spanning 80+ fine-grained risk types across multimodal QA, adversarial attack, and dynamic-rule evaluation settings, including cross-modal joint-risk cases where each modality is harmless in isolation but their composition implies unsafe intent. Across six benchmark families (35 datasets), SingGuard achieves state-of-the-art average F1 in every family. Dynamic-rule evaluation further shows improved policy-following accuracy from 0.6465 to 0.7415 under runtime policy shifts. Our code is available at https://github.com/inclusionAI/Sing-Guard.

SingGuard Team

1AI Security Lab, Ant Group

Language Model Safety Evaluation LLM Guardrails Multimodal Large Language Models Multimodal Model Evaluation

Jun 21, 2026cs.CV

Catching Lies Without Sending the Video: Privacy-Preserving Multimodal Deception Detection

Frontier multimodal models can guess whether a person is lying from a testimony video. To do so, they stream that raw face and voice to a third-party model. We ask whether the heavy media is needed at all. On the Real-life Trial Deception dataset, Whissle on-device speech and vision stack extracts a compact digest: transcript, emotion, age, gender, intent distributions, a deception intent filter, fluency and rhythm, per-frame facial behaviour, and prosody. Under speaker-independent evaluation, we report three findings. A small classifier on this digest reaches AUC 0.741, matching Gemini 2.5 Pro on full video. Handing the digest to a frontier LLM reaches AUC 0.755 with Claude Opus 4.8 at 7.8X fewer input tokens, with no media leaving the device. The reported 75% accuracy is a speaker-leakage artifact. We release code and experiments.

Nikita Sharma, Pranav Sara, Karan Singla

University of California, San Diego · Case Western Reserve University, Ohio · WhissleAI, USA

Privacy-Preserving ML Multimodal Model Evaluation

Jun 21, 2026cs.CL

Look Light, Think Heavy: What Multimodal Chain-of-Thought Reasoning Can and Cannot Do

Chain-of-Thought (CoT) has become a standard method for improving reasoning capabilities in large language models (LLMs) by eliciting step-by-step thinking, but its effectiveness in multimodal tasks remains unclear. In this paper, we aim to systematically investigate the key question: What can multimodal Chain-of-Thought reasoning do, and where and why does it fall short? To this end, we evaluate 12 multimodal tasks across perception and reasoning categories using both 14 non-reasoning models and 8 reasoning models. Our analysis reveals several important findings: (1) CoT is not a free lunch and should be used selectively depending on the specific requirements of each task. For perception tasks, CoT can lead to undesirable side effects, such as reduced performance in visual grounding and object counting. In contrast, it proves effective for reasoning tasks involving mathematical, scientific, and multi-image reasoning; (2) Compared to original models, existing open-source multimodal reasoning models often yield only marginal overall improvements, possibly due to an overemphasis on mathematical reasoning at the expense of broader capabilities; (3) Visual reasoning remains a key bottleneck for current multimodal CoT, as models exhibit a Look Light, Think Heavy pattern where verbal reflection rises and falls during reasoning, whereas visual reflection consistently diminishes. These findings suggest that while multimodal CoT handles verbal reflection relatively well, it lacks the ability to maintain deep visual introspection throughout the reasoning process.

Zhuoran Jin, Kejian Zhu, Hongbang Yuan +5

The Laboratory of Cognition and Decision Intelligence for Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing, China · School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China

Multimodal Reasoning VLM Reasoning Multimodal CoT Reasoning Multimodal Model Evaluation Multimodal Understanding

Jun 21, 2026cs.AI

Efficient Multimodal Clinical Question Answering for Pulmonary Embolism Risk Assessment

Pulmonary embolism (PE) is a high risk cardiopulmonary condition whose management requires both timely diagnosis and reliable assessment of future clinical risk. Because PE care routinely combines computed tomography pulmonary angiography (CTPA), radiology interpretation, and longitudinal electronic health record (EHR) evidence, it provides a clinically meaningful setting for evaluating compact multimodal language models. In this work, we build a benchmark using efficient multimodal large language models (MLLMs) on INSPECT, a multimodal PE dataset containing 23,248 CTPA studies from 19,402 patients. We formulate eight diagnostic and prognostic tasks as structured clinical question answering problems and evaluate on typical efficient MLLMs under CTPA-Only, EHR-Only, and CTPA+EHR settings with zero-shot and few-shot prompting. Results show that Gemma4 E4B and Gemma4 E2B perform more strongly when EHR evidence is available, especially under CTPA+EHR input. Task level analysis further shows that PE diagnosis achieves higher performance than prognostic tasks, particularly readmission prediction. These observations suggest that compact multimodal models have the great potential in early stage PE risk detection and explanation.

Xiangyuan Xue, Yang Yu, Yan Gao +5

University of Auckland, Auckland, New Zealand · University of Cambridge, Cambridge, United Kingdom · University of Adelaide, Adelaide, Australia +1

Efficient Multimodal Inference Multimodal QA Medical VLMs Multimodal Model Evaluation Clinical Risk Prediction

Jun 21, 2026cs.CV

T-IMPACT: A Severity-Aware Benchmark for Contextual Image-Text Manipulation

Recent advances in vision-language models and generative editing systems have made it increasingly easy to produce persuasive multimodal misinformation by altering images, text, or both jointly. However, existing datasets focus mainly on authenticity, out-of-context mismatch, or manipulation type, and rarely capture how strongly an edit changes the likely interpretation of a post. We introduce T-IMPACT, a first-release severity-aware benchmark for manipulated news-style image-text pairs. T-IMPACT contains 98,786 examples spanning pristine, image-only, text-only, and joint manipulations, with a calibrated continuous severity signal, coarse low/medium/high labels, and supporting grounding metadata. Starting from a news image-text pair, the pipeline extracts semantic anchors, grounds them spatially, performs localized image edits and constrained caption rewrites, and calibrates contextual-impact scores using limited human ratings. In this release, the calibrated continuous score is the primary severity target, while the low/medium/high bands should be interpreted as coarse operating buckets rather than balanced classes. Experiments show that current models recover some authenticity signal, but severity prediction remains substantially harder and only weakly aligned with human judgment. T-IMPACT provides an initial benchmark for studying multimodal manipulation beyond binary real/fake classification toward graded contextual impact.

Gagandeep Singh, Aaditya Yadav, Priyanka Singh

The University of Queensland · Brisbane, Australia

Multimodal Model Evaluation

Jun 20, 2026cs.LG

Bayesian Adaptation Gym: A Benchmark for the Bayesian Low-Rank Adaptation of Multi-Modal Language Models

Large multi-modal language models are increasingly deployed in high-stakes domains, making well-calibrated uncertainty essential. Traditional Bayesian methods approximate posteriors over all model weights, which becomes intractable for modern large models. For this reason, recent work instead considers Bayesian low-rank adaptation to enable tractable posterior approximation. Due to a lack of a standardized benchmark to evaluate these approaches, it remains unclear where these methods provide meaningful benefits. To fill this gap, we introduce Bayesian Adaptation Gym (BAG), a benchmark for the Bayesian adaptation of multi-modal language models. BAG provides reference implementations of classic Bayesian baselines and state-of-the-art adaptation methods, along with a multi-modal dataset and task suite designed to probe calibration, robustness under distribution shift, and decision-making under uncertainty via active learning. Using BAG, we conduct and report extensive experiments across model sizes, datasets, and tasks to highlight the successes and failures of current Bayesian adaptation approaches. To enable further research, BAG is fully open source: https://github.com/SRI-CSL/BayesAdapt.

Colin Samplawski, Ramneet Kaur, Manoj Acharya +2

Neuro-Symbolic Computing and Intelligence Research Group, Computer Science Laboratory, SRI International

Distribution Shift Robustness Multimodal Large Language Models Low-Rank Adaptation Language Model Calibration Multimodal Model Evaluation Bayesian LoRA

Jun 20, 2026cs.CV

SAGE: An Expert-Annotated South Asian GI Endoscopy Dataset for Multimodal Learning and Hallucination Analysis

Gastrointestinal cancers represent a growing health burden in the South Asian region, driven largely by rapid changes in socio-economic conditions & lifestyle habits. However, early diagnosis of such malignancies remains a significant challenge, largely due to a lack of modern equipment, lack of financial support, and a scarcity of GI experts. AI-assisted diagnosis & report generation, show great promise in alleviating this problem by providing low-skill manpower the technical expertise to perform diagnosis. However, almost all open-source, publicly available datasets are predominantly collected from the European region, with no representation from the South Asian region. The lack of open-source GI datasets from diverse geographic regions has made it difficult to assess whether population bias is present in existing models, and to develop geographically inclusive AI tools for automated GI diagnosis. To address this gap, we introduce SAGE: An Expert-Annotated South Asian GI Endoscopy dataset for image captioning, multi-label classification, and visual question answering (VQA) tasks. It consists of 1,300 images, their captions along with hallucination tag, 18 labels and 14,726 question-answer pairs making it well-suited for diverse range of tasks including classification, benchmarking, and fine-tuning large multimodal models (LMMs). We further conducted benchmarking of multi-class classifiers on the effect of population shift in GI imaging AI tasks, and contemporary LMMs on their performance. Our study reveals that task-specific models, such as multi-class classification models, suffer the most, with an average performance drop of 58% when evaluated on the South Asian dataset. For contemporary LMMs, benchmarking reveals a substantial drop in the average GREEN score for anatomical landmark detection (0.308) and abnormality detection (0.410).

Niyoj Oli, Sachin Acharya, Sandesh Pokhrel +7

Nepal Applied Mathematics and Informatics Institute for Research, Nepal · University of Utah, USA · GastroIntestinal Department, Dhulikhel Hospital, Nepal +3

Endoscopy Medical Imaging Multimodal Hallucination Medical VLMs Multimodal Model Evaluation Medical Image Benchmarks+1

Jun 19, 2026cs.CV

Chehre: An Emoji-Prompted Dataset to Explore Perceptual Flexibility in Video Language Models

Do people perceive the same facial expression in the same way? Should we expect vision models to be flexible in how they perceive facial expressions? Facial expressions are nonverbal social signals used in human interaction, but facial expression recognition datasets often focus on a single deterministic annotation per sample. We introduce Chehre, an emoji-prompted video dataset with a wide range of dynamic facial expressions for exploring perceptual variation. In Chehre, 203 participants were prompted to express and record 40 facial emojis. Later, their facial motions were transferred onto synthetic faces to preserve privacy. A separate group annotated the videos, resulting in 2,111 videos annotated by 1,242 perceivers, with ~30 annotators per video. Chehre enables us to define a new task: "distributional expression recognition", which tests whether a model can reproduce the variation observed across annotator responses. We test a selection of video language models on our task. Interestingly, we find that persona prompting can act as a controllable way to shift model perception while helping models better capture the variation observed across human annotators. The dataset and code are available at https://chehre-dataset.github.io/.

Bita Azari, Zoe Stanley, Avneet Batra +4

Simon Fraser University, Canada

VLM Evaluation Video-Language Model Evaluation Facial Expression Recognition Persona Prompting Multimodal Model Evaluation Prompt Engineering

Jun 19, 2026cs.CL

CulMind: Benchmarking Multimodal Understanding and Reasoning in Chinese Cultural Heritage

Evaluating Multimodal Large Language Models (MLLMs) in Chinese Cultural Heritage (CCH) requires fine-grained reasoning over visual, textual, stylistic, and historical clues. However, existing CCH benchmarks mainly emphasize final-answer accuracy, while the accuracy and completeness of reasoning processes remain underexplored. To address this gap, we introduce CulMind and CulMind-R: a high-quality benchmark for multimodal CCH covering 50 tasks from collections of more than 100 museums, and a 24-task reasoning subset that adaptively defines task-specific dimensions for reasoning process evaluation. To evaluate reasoning quality, we propose ReaScore, a task-adaptive metric that evaluates reasoning by automatically weighting task-relevant dimensions. Experiments on 14 leading MLLMs reveal a substantial gap between answers and reasoning, especially on challenging tasks. Further analysis shows that task-adaptive dimension selection and weighting better align evaluation results with expert judgments. Overall, our benchmark and metric support a more expert-aligned assessment of CCH understanding and offer a transferable reference for broader evaluations of cultural heritage. We publicly release the data, code, and evaluation scripts at https://github.com/ZevTsao/CulMind to facilitate reproducible research.

Zhangwei Cao, Shuhan Fan, Yuting Wei +5

University of International Relations · 2Kedge Business School · 3Peking University +3

Benchmark Construction Reasoning Evaluation Multimodal Reasoning Cultural Knowledge in Language Models Cultural Heritage Multimodal Model Evaluation+1

Jun 19, 2026cs.CV

Cross-Modal Corroboration for Annotation-Free Wildlife Monitoring

Scaling wildlife monitoring for real-world conservation deployments requires automated analysis of smart sensors that operate under severe annotation scarcity. We propose leveraging expert knowledge of species activity patterns as an annotation-free validation signal for multimodal monitoring pipelines. We operationalize agreement as the alignment of independently derived hourly activity curves both with each other and with published behavioral priors-a three-way convergence that rules out shared-data confounds and dataset-internal correlation as alternative explanations. Our vision pipeline combines zero-shot species detection via BioCLIP 2, sliced inference to handle deployment-constrained camera positioning, and geometry-based geographic localization from camera trap imagery. Our acoustic pipeline detects species vocalizations via a fine-tuned classifier. We validate the pipeline on a breeding herd of Milu deer and demonstrate that both modalities independently recover activity patterns consistent with known deer behavioral ecology with minimal manual annotation. The framework applies to species detectable in both visual and acoustic modalities for which behavioral priors are documented in the literature, suggesting a practical path toward self-validating wildlife-monitoring pipelines at conservation scale.

Bharath Pillai, Varun Viswapriyan, Christopher Stewart +2

The Ohio State University

Animal Behavior Analysis Multimodal Model Evaluation Wildlife Monitoring

Jun 18, 2026cs.CL

StylisticBias: A Few Human Visual Cues Drive Most Social Biases in MLLMs

Multimodal large language models (MLLMs) are increasingly deployed in personally and societally consequential settings, yet the visual cues that shape how these models judge people remain poorly understood. Prior work often compares different (groups of) individuals, making it difficult to separate appearance effects from identity differences. We introduce StylisticBias, a controlled benchmark for evaluating attribute-level social bias in MLLMs. We generate 500 photorealistic base faces and create about 50 single-attribute variations per face, producing about 25K images. This design keeps identity fixed and changes one visual attribute at a time. It lets us measure how specific cues shift model judgments. We evaluate six MLLMs across 25 binary social judgment scenarios. We find that age and body type dominate identity-level effects, while fashion style and other visual cues drive the largest attribute-level shifts. We further find that about 15 attributes account for nearly 80% of the total variation, showing that bias is concentrated in a small set of visual cues. Sensitivity is strongest in judgments that are semantically aligned with appearance, especially socioeconomic and style-related judgments. We release StylisticBias as a benchmark for fine-grained bias evaluation in multimodal models. Code and dataset: https://github.com/timo-cavelius/StylisticBias and https://hf.co/datasets/shaghayegh/stylistic-bias-dataset.

Shaghayegh Kolli, Timo Cavelius, Nafiseh Nikeghbal +2

Technical University of Munich · Munich Center for Machine Learning · Princeton Center for Information and Technology Policy

Social Bias in Language Models Multimodal Large Language Models Demographic Bias in Language Models Multimodal Model Evaluation

Jun 18, 2026cs.CV

ROSE: Benchmarking the Perception-to-Action Gap in Multimodal Models

Multimodal large language models (MLLMs) are increasingly expected to act on visual information, yet the same scene may require different actions under different task contexts. How reliably can a model turn the same visual evidence into the action required by the current context? To answer this question, we introduce \textsc{ROSE} (\textbf{R}eference-conditioned \textbf{O}ddity and \textbf{S}ymbolic \textbf{E}xecution), a controlled benchmark that holds the visual scene fixed while varying region constraints and required symbolic outputs. Through coupled counting and coordinate-action tasks, \textsc{ROSE} tests whether models can infer an implicit majority reference and act on the resulting fine-grained visual evidence under changing contexts. Across nine recent MLLMs, performance drops by as much as 44.5 percentage points from counting-oriented tasks to region-conditioned action, despite 98.8% human performance. The gap persists on paired scenes and regions for which the same model returns the correct count, while global-click and matched local controls show that coordinate grounding explains only part of the loss, revealing a distinct, model-dependent bottleneck in turning shared visual evidence into context-specific actions.

Yihao Wang, Zijian He, Jie Ren +1

Sun Yat-sen University · Shaanxi Normal University

Multimodal Grounding Multimodal Reasoning Visually Grounded Reasoning Multimodal Model Evaluation Multimodal Understanding

Jun 17, 2026cs.CV

Beyond the Current Observation: Evaluating Multimodal Large Language Models in Controllable Non-Markov Games

Deploying multimodal foundation models as closed-loop policies increasingly requires conditioning actions on observations that are no longer visible. However, existing benchmarks either expose the full state, conflate hidden-state reconstruction with other agent skills, or test recall only after an episode has ended. We introduce RNG-Bench (Reconstructive Non-Markov Games), a benchmark suite designed to isolate a base model's ability to reconstruct past observations and act on them during multi-step interaction. RNG-Bench includes two complementary games: Matching Pairs, where card identities briefly revealed at specific locations must later be recalled, and 3D Maze, where egocentric views must be integrated into a spatial map. Both games are evaluated under a unified harness with three controlled difficulty axes: grid size, visual pattern, and observation modality. The benchmark further introduces a head-to-head duel protocol to control for instance-level variance and a Memory Gap metric that disentangles forgetting from poor action selection. The hardest configurations require contexts of roughly 128K tokens and 350 image inputs per episode, and remain far from saturated by frontier MLLMs. Memory Gap analysis shows that most residual errors stem from forgetting earlier observations rather than from suboptimal decision making. Finally, fine-tuning Qwen3.5-9B on optimal-policy rollouts and filtered model demonstrations improves performance on RNG-Bench and transfers to existing benchmarks without degrading general multimodal capability.

Shengyuan Ding, Xilin Wei, Xinyu Fang +4

1Fudan University · 2Shanghai Innovation Institute · 3Shanghai Artificial Intelligence Laboratory +2

Multimodal Memory Sequential Decision Making AI Agent Benchmarks Multimodal Foundation Models Multimodal Agents Multimodal Model Evaluation+1

Jun 16, 2026cs.CV

Million-scale multimodal pollen microscopy with expert-guided foundation models

Automated pollen identification from microscopy remains a bottleneck in aerobiology, palaeoecology and biodiversity monitoring, because scalable systems must generalise across specimen preparation, scanner settings and geographic origins while retaining palynological interpretability. To address this gap, we present a million-scale multimodal pollen microscopy resource, Pollen AI Atlas, assembled from pure-species whole-slide bright-field images spanning four geographic origins, four scanner settings and 46 taxon labels across 31 botanical families. Seeded by one manually selected exemplar per source slide, token-level mining and filtering produced 1,511,390 released grain detections with 99.6% proposal precision in expert-curated test regions. Each detection was paired with machine-generated grain-level morphological captions from five open-weight vision-language models, guided by expert-verified palynological anchors, yielding structured descriptions of aperture systems, wall ornamentation, shape and size. Among the evaluated models, Gemma4 provided the most controlled primary caption set, combining tight length control, no leakage and the strongest text-retrieval performance. Baseline benchmarks with frozen visual features reached 88.16% top-1 accuracy, while cross-regional retrieval showed that caption-derived text embeddings remained robust when image similarity degraded (mAP@20 0.811 versus 0.262). Released data, annotations, captions, splits, code, and weights provide a benchmark for pollen recognition, cross-regional domain adaptation and domain-specific multimodal microscopy learning.

András Biricz, Björn Gedda, Donát Magyar +4

Department of Physics of Complex Systems, ELTE Eötvös Lor´and University, Budapest, Hungary · 2The Palynological Laboratory at the Swedish Museum of Natural History, Stockholm, Sweden · 3National Centre for Public Health and Pharmacy, Budapest, Hungary +4

Domain Adaptation Multimodal Model Evaluation Microscopy Image Analysis

Jun 16, 2026cs.CL

The Slop Paradox: How Synthetic Standardization Erodes Clinical Uncertainty and Cross-Modal Alignment in AI-Rewritten Radiology Reports

AI-assisted clinical documentation tools increasingly summarize, standardize, and reformat radiology reports using large language models (LLMs). We present a controlled measurement of the resulting information degradation. Using 450 chest X-ray reports from the Indiana University dataset, we generate synthetic versions via three realistic LLM rewriting tasks: EHR summarization, standardized rewriting, and teaching case preparation. We measure entity erosion (via medical NER), hedging collapse (loss of clinical uncertainty language), and cross-modal alignment degradation (via BiomedCLIP image-text similarity). Our central finding is a dissociation between information loss and cross-modal fidelity. EHR summarization is the most destructive at the content level, eroding 51.4% of clinical entities and 43.7% of hedging language, yet it preserves image-text alignment almost entirely (a 2.5% drop). The two tasks meant to produce cleaner training data, standardized rewriting and teaching case preparation, do the reverse: they preserve more entities (26.8% and 29.3% eroded) but cause 14.9-16.5% alignment drops, six to seven times those of EHR summarization. We term this the slop paradox: rewriting that makes clinical text look cleaner for multimodal training is precisely what pulls it away from the image. Contrary to our pre-specified hypothesis, rare pathologies were not preferentially degraded: across nine rare-versus-common comparisons, no difference survived multiple-comparison correction, and nominal differences ran in the opposite direction (common > rare), so contamination is invisible to condition-specific monitoring. The dominant determinant of degradation is the type of AI rewriting task, not the clinical content. These findings bear on multimodal medical AI dataset construction and the governance of AI-assisted clinical documentation.

Samar Ansari

School of Computing and Engineering Sciences University of Chester Chester, CH1 4BJ, United Kingdom

Medical VLMs Vision-Language Alignment Multimodal Model Evaluation

Jun 16, 2026cs.CL

MODE-RAG: Manifold Outlier Diagnosis and Energy-based Retrieval-Augmented Generation Evaluation

While Multimodal Retrieval-Augmented Generation (M-RAG) enhances Large Vision-Language Models, it remains highly susceptible to cross-modal hallucinations, causal fabrications, and sycophancy. Furthermore, existing mitigation pipelines often face an intervention paradox: static rules tend to unnecessarily disrupt accurate generations, whereas leaving the multi-modal reasoning completely unguided allows existing mismatches to cascade into severe logical fabrications. To quantify and mitigate these hallucinations, we propose a Multi-Agent system, MODE-RAG, driven by Variational Free Energy (VFE) and internal attention states to dynamically gate interventions. High-risk queries are routed to five stage-specific agents, integrating Monte Carlo Tree Search (MCTS) for rigorous causal derivation and logit perturbations to penalize sycophancy. Dedicated Correction and Overseer agents ensure formatting stability and perform post-hoc factual verification. To objectively evaluate our approach, we introduce ModeVent, a challenging subset derived from the MultiVent dataset. Extensive experiments indicate that our system effectively reduces hallucination rates and logical fabrication, significantly improving the robustness of M-RAG systems.

Zehang Wei, Jiaxin Dai, Jiamin Yan +1

School of AI and Automation, Huazhong University of Science and Technology, China · School of Computer Science & Tech, Huazhong University of Science and Technology

Multimodal RAG Multi-Agent LLM Systems Multimodal Hallucination Multimodal Model Evaluation VLM Hallucination Mitigation

Jun 15, 2026cs.LG

Multimodal Evaluator Preference Collapse: Cross-Modal Coupling in Self-Evolving Agents

When AI agents use language models to evaluate their own outputs in a feedback loop, systematic biases emerge. We show that Evaluator Preference Collapse (EPC) is dramatically amplified in multimodal settings. Using GPT-4o to evaluate DeepSeek-chat across text and visual tasks, we find that a single strategy (step_by_step) absorbs 48.4% of all weight -- 3.2x the collapse observed in text-only self-evaluation -- while three visual-domain strategies receive only 9.1% combined weight. We then demonstrate a novel phenomenon we term cross-modal coupling: evaluator preferences acquired on one modality transfer to and corrupt strategy selection on another. Through a four-phase isolation training paradigm, we measure coupling coefficients and document strategy inversion -- the optimal strategy for a modality reverses after cross-modal exposure. A Phase 3 statistical validation across five evaluator configurations (N=80 total independent repetitions, ~~35,000 API calls) with both text-proxy and real-image visual tasks finds: cross-model evaluation produces strong coupling (JSD~~0.19-0.34), real-image inputs yield the most directionally consistent signal (mean gamma_{T->V}=1.145, gamma_{V->T}=0.937, 70% T->V, Cohen's d=0.56), and self-evaluation provides near-complete immunity -- 97% of runs (N=30) yield zero coupling (JSD=0.003, d=0.07). Three methodological ablations and multi-executor validation confirm the effect is not a structural artifact. We introduce the coupling matrix indexed by evaluator identity, release the MM-EPC framework, and identify cross-model evaluator architecture as the primary risk factor for preference drift. Code and data: https://github.com/aidless/mm-epc.

Zewen Liu

Qilu Institute of Technology, School of Software Engineering · Tai’an, Shandong, China

AI Agent Evaluation Multimodal Model Evaluation

Jun 14, 2026cs.CV

Mind the Gap: Diagnosing Constraint Discovery Failures in Text-in-Image Editing

A key challenge in multimodal reasoning is determining which visual dependencies become relevant under a specific task, rather than merely recognizing visible content. We study this through edit-induced constraint discovery in text-in-image editing, a controlled diagnostic setting where a local text change can activate secondary consistency constraints: given a valid editing instruction and an image, can a model identify the secondary regions that must also change? Across 461 diagnostic cases, four MLLMs, and 19 constraint subtypes, models recover only 46% case-level macro recall under unguided prompting versus 94% when constraints are explicitly provided, suggesting that a substantial portion of the failure arises when models must decide which unstated dependencies to surface. Oracle-field decomposition shows that case-specific causal explanations are the most effective partial guidance (0.782 recall), above region names (0.610) or type labels (0.646), suggesting that edit-specific causal cues account for much of the oracle gain. A downstream experiment further shows that higher self-discovery recall does not necessarily improve task performance: unverified self-discovery introduces false positives that offset recall gains, motivating precision-aware constraint elicitation.

Rui Gui

Central South University

Text-Guided Image Editing Multimodal Large Language Models Multimodal Reasoning Multimodal Model Evaluation Multimodal Understanding

Jun 14, 2026cs.AI

UrbanWell: Benchmarking Multimodal Large Language Models for Spatio-Temporal Urban Wellbeing Analytics

Understanding urban wellbeing from multimodal data requires integrating heterogeneous spatial and temporal signals, posing significant challenges for current multimodal large language models (MLLMs). We introduce UrbanWell, a large-scale benchmark designed to systematically evaluate the spatio-temporal reasoning capabilities of MLLMs for urban wellbeing analytics through joint modeling of satellite and street view imagery. UrbanWell spans 38 cities across multiple years and includes diverse indicators covering (1) environmental conditions (CO

_2

, NO

_2

, PM

{2.5}

, and Normalized Difference Vegetation Index), (2) spatial accessibility (minimum distance to supermarkets and restaurants), (3) urban form (road length, road density, and land use), (4) urban vitality (population, economic activity diversity, and land use diversity), and (5) subjective perception attributes (e.g., safety, beauty, liveliness, wealth, and quietness). All indicators are aligned at grid level to enable standardized evaluation. Beyond static prediction, UrbanWell defines temporal reasoning tasks, including future value forecasting from historical observations and temporal trend classification. We benchmark 15 state-of-the-art representative MLLMs in a zero-shot setting, providing a comprehensive comparative evaluation across spatial and temporal dimensions. Experimental results indicate that while MLLMs capture salient spatial and perceptual cues, their performance varies substantially across heterogeneous urban indicators spanning environment and subjective perception. UrbanWell serves as a unified benchmark for evaluating multimodal spatial and temporal reasoning in urban wellbeing analytics, offering a standardized testbed for systematic assessment and future research on multimodal urban intelligence. Our codes and datasets are accessible via https://github.com/axin1301/UrbanWell-Benchmark.

Yanxin Xi, Xiang Su, Jie Feng +3

Department of Computer Science, University of Helsinki, Helsinki, Finland · Department of Computer Science, Department of Agricultural Sciences, University of Helsinki, Helsinki, Finland · Zhongguancun Academy, Beijing, China +2

Spatiotemporal Reasoning Multimodal Large Language Models Spatiotemporal Forecasting Multimodal Model Evaluation

Jun 14, 2026cs.CV

OmniTraffic: A Controllable Generation Pipeline and Benchmark for Spatio-Temporal Traffic Reasoning

Traffic scene understanding requires models to reason beyond object recognition, including lane topology, multi-view geometry, temporal evolution, and signal-phase semantics. However, existing traffic-oriented multimodal benchmarks largely emphasize passive visual recognition or isolated video understanding, offering limited support for evaluating structure-aware traffic reasoning under controlled conditions. We introduce OmniTraffic, a controllable generation pipeline and benchmark for spatio-temporal traffic reasoning. Built around 12 real-world intersections reconstructed into editable 3D traffic environments and complemented by surveillance footage from two countries, OmniTraffic supports both controlled and natural-condition evaluation. It defines a three-level task hierarchy spanning scene perception, multi-view and temporal reasoning, and decision support. Using structured traffic metadata, OmniTraffic generates synchronized multi-view VQA samples covering vehicle states, lane functions, view--BEV correspondence, temporal dynamics, and signal-phase analysis, resulting in 8M VQA samples and a 3K human-verified test set. Evaluation of eleven frontier MLLMs reveals a large human--model gap, with the most pronounced failures in topology-grounded and spatio-temporal reasoning tasks. Fine-tuning a lightweight MLLM on simulated OmniTraffic data further improves performance on real-world traffic scenes, demonstrating the value of simulation-generated supervision for traffic-specific multimodal reasoning. Beyond a fixed dataset, OmniTraffic provides an extensible pipeline with configurable intersections, camera views, traffic demands, signal phases, visual conditions, and rare events.

Maonan Wang, Zhengyan Huang, Kemou Jiang +13

The Chinese University of Hong Kong, Shenzhen · Shanghai AI Lab · Beihang University +3

Autonomous Driving Datasets Spatiotemporal Reasoning Autonomous Driving Perception Multimodal Reasoning Multimodal Model Evaluation

Jun 14, 2026cs.CV

On the Adversarial Robustness of Multimodal LLM Judges

Multimodal Large Language Models (MLLMs) are increasingly used as automated judges, e.g., for image quality and safety assessment. However, their adversarial robustness remains largely unexplored, threatening the fairness and reliability of automated judging. To bridge this gap, we introduce RobustMLLMJudge, the first general framework for evaluating the adversarial robustness of general-purpose MLLMs when functioning as judges. It covers diverse attacks against popular judge approaches across quality and safety evaluation scenarios. Using RobustMLLMJudge, we reveal that i) different MLLM judges are highly vulnerable to score-inflating adversarial attacks; and ii) although effective, these attack methods face a critical challenge due to unique constraints in the evaluation protocols of MLLM judges. We further propose MGSIA, namely Manifold-Guided Semantic Induction Attack, a novel method that bypasses these constraints to enable more effective and transferable attacks on MLLM judges. The core idea of MGSIA is to combine affirmative semantic induction with high-score manifold alignment: it maximizes the probability that judges yield affirmative responses (e.g., "Yes") to binary semantic queries, while regularizing adversarial representations toward high-score centers estimated from proxy protocols. Together, these objectives yield transferable score-inflating perturbations. Extensive experiments demonstrate the superiority and generalizability of MGSIA in deceiving advanced MLLM judges under different evaluation scenarios, highlighting the need for robust MLLM judges. Code and data will be made available at https://github.com/mala-lab/RobustMLLMJudge.

Zihan Wang, Guansong Pang, Zelin Liu +3

School of Computer Science and Engineering, Beihang University, Beijing, China · School of Computing and Information Systems, Singapore Management University, Singapore · State Key Laboratory of Virtual Reality Technology and System, Beihang University, Beijing, China +1

LLM-as-a-Judge Adversarial Robustness Multimodal Model Evaluation