VLM Evaluation

VLM: Vision-Language Model

Follow this topic Ask about this topic

Vision-language model evaluation studies what these systems can infer from images and video, and whether their answers are supported by visual evidence. Researchers use broad benchmarks alongside controlled probes, counterfactual edits, and stage-wise tests covering tasks such as spatial reasoning, charts and documents, medical imaging, and long-video understanding. They assess not only accuracy but also grounding, hallucination, calibration, robustness, bias, and abstention, often comparing models with human or specialist judgments. A central challenge is that high benchmark scores can reflect language priors or shortcuts rather than reliable visual understanding, and performance often falters on fine-grained, temporal, compositional, or unfamiliar tasks.

Latest papers

999 papers122 papers in the last four weeks, up 12% on the four weeks before. 0.7% of all new papers.

Mar 26, 2026

RealChart2Code: Advancing Chart-to-Code Generation with Real Data and Multi-Task Evaluation

Vision-Language Models (VLMs) have demonstrated impressive capabilities in code generation across various domains. However, their ability to replicate complex, multi-panel visualizations from real-world data remains largely unassessed. To address this gap, we introduce RealChart2Code, a new large-scale benchmark with over 2,800 instances grounded in authentic datasets and featuring tasks with clear analytical intent. Crucially, it is the first benchmark to systematically evaluate chart generation from large-scale raw data and assess iterative code refinement in a multi-turn conversational setting. Our comprehensive evaluation of 14 leading VLMs on RealChart2Code reveals significant performance degradation compared to simpler benchmarks, highlighting their struggles with complex plot structures and authentic data. Our analysis uncovers a substantial performance gap between proprietary and open-weight models and confirms that even state-of-the-art VLMs often fail to accurately replicate intricate, multi-panel charts. These findings provide valuable insights into the current limitations of VLMs and guide future research directions. We release the benchmark and code at https://github.com/Speakn0w/RealChart2Code.

Jiajun Zhang, Yuying Li, Zhixun Li +13

USTC · CASIA · THU +6

Chart-to-Code Generation Code Generation Code Generation Benchmarks VLM Evaluation

Mar 26, 2026

Humans vs Vision-Language Models: A Unified Measure of Narrative Coherence

We study narrative coherence in visually grounded stories by comparing human-written narratives with those generated by vision-language models (VLMs) on the Visual Writing Prompts corpus. Using a set of metrics that capture different aspects of narrative coherence, including coreference, discourse relation types, topic continuity, character persistence, and multimodal character grounding, we compute a narrative coherence score. We find that VLMs show broadly similar coherence profiles that differ systematically from those of humans. In addition, differences for individual measures are often subtle, but they become clearer when considered jointly. Overall, our results indicate that, despite human-like surface fluency, model narratives exhibit systematic differences from those of humans in how they organise discourse across a visually grounded story. Our code is available at https://github.com/GU-CLASP/coherence-driven-humans.

Nikolai Ilinykh, Hyewon Jang, Shalom Lappin +2

Dept. of Philosophy, Linguistics, and Theory of Science, University of Gothenburg · School of Electronic Engineering and Computer Science, Queen Mary University of London · Dept. of Informatics, King’s College London

Visual Storytelling VLM Evaluation

Mar 26, 2026

THEMIS: Towards Holistic Evaluation of MLLMs for Scientific Paper Fraud Forensics

We present THEMIS, a novel multi-task benchmark designed to comprehensively evaluate multimodal large language models (MLLMs) on visual fraud reasoning within real-world academic scenarios. Compared to existing benchmarks, THEMIS introduces three major advances. (1) Real-World Scenarios and Complexity: Our benchmark comprises over 4,000 questions spanning seven scenarios, derived from authentic retracted-paper cases and carefully curated multimodal synthetic data. With 60.47% complex-texture images, THEMIS bridges the critical gap between existing benchmarks and the complexity of real-world academic fraud. (2) Fraud-Type Diversity and Granularity: THEMIS systematically covers five challenging fraud types and introduces 16 fine-grained manipulation operations. On average, each sample undergoes multiple stacked manipulation operations, with the diversity and difficulty of these manipulations demanding a high level of visual fraud reasoning from the models. (3) Multi-Dimensional Capability Evaluation: We establish a mapping from fraud types to five core visual fraud reasoning capabilities, thereby enabling an evaluation that reveals the distinct strengths and specific weaknesses of different models across these core capabilities. Experiments on 16 leading MLLMs show that even the best-performing model, GPT-5, achieves an overall performance of only 56.15%, demonstrating that our benchmark presents a stringent test. We expect THEMIS to advance the development of MLLMs for complex, real-world fraud reasoning tasks.

Tzu-Yen Ma, Bo Zhang, Zichen Tang +20

Beijing University of Posts and Telecommunications

VLM Evaluation Image Forgery Detection

Mar 26, 2026

Mechanistically Interpreting Compression in Vision-Language Models

Compressed vision-language models (VLMs) are widely used to reduce memory and compute costs, making them a suitable choice for real-world deployment. However, compressing these models raises concerns about whether internal computations and safety behaviors are preserved. In this work, we use causal circuit analysis and crosscoder-based feature comparisons to examine how pruning and quantization fundamentally change the internals across representative VLMs. We observe that pruning generally keeps circuit structure intact but rotates and attenuates internal features, while quantization modifies the circuits at a higher level yet leaves the surviving features better aligned. Leveraging this insight, we also introduce VLMSafe-420, a novel benchmark that pairs harmful inputs with matched benign counterfactuals across various safety categories. Our findings show that pruning causes a sharp drop in genuine refusal behavior, suggesting that the choice of compression has safety implications.

Veeraraju Elluru, Arth Singh, Roberto Aguero +3

Indian Institute of Technology Jodhpur · AIM Intelligence · National Institute of Technology Agartala +3

Cultural Safety Evaluation of VLMs LVLM Pruning VLM Quantization VLM Interpretability Neural Network Pruning VLM Evaluation

Mar 25, 2026

How Far Are Vision-Language Models from Constructing the Real World? A Benchmark for Physical Generative Reasoning

The physical world is not merely visual; it is governed by rigorous structural and procedural constraints. Yet, the evaluation of vision-language models (VLMs) remains heavily skewed toward perceptual realism, prioritizing the generation of visually plausible 3D layouts, shapes, and appearances. Current benchmarks rarely test whether models grasp the step-by-step processes and physical dependencies required to actually build these artifacts, a capability essential for automating design-to-construction pipelines. To address this, we introduce DreamHouse, a novel benchmark for physical generative reasoning: the capacity to synthesize artifacts that concurrently satisfy geometric, structural, constructability, and code-compliance constraints. We ground this benchmark in residential timber-frame construction, a domain with fully codified engineering standards and objectively verifiable correctness. We curate over 26,000 structures spanning 13 architectural styles, ach verified to construction-document standards (LOD 350) and develop a deterministic 10-test structural validation framework. Unlike static benchmarks that assess only final outputs, DreamHouse supports iterative agentic interaction. Models observe intermediate build states, generate construction actions, and receive structured environmental feedback, enabling a fine-grained evaluation of planning, structural reasoning, and self-correction. Extensive experiments with state-of-the-art VLMs reveal substantial capability gaps that are largely invisible on existing leaderboards. These findings establish physical validity as a critical evaluation axis orthogonal to visual realism, highlighting physical generative reasoning as a distinct and underdeveloped frontier in multimodal intelligence. Available at https://luluyuyuyang.github.io/dreamhouse

Luyu Yang, Yutong Dai, An Yan +3

Salesforce AI Research

Mar 25, 2026

Evaluating Vision-Language Models for Image Quality Assessment using Psychophysical Data

Psychophysical experiments remain the most reliable approach for perceptual image quality assessment (IQA), yet their cost and limited scalability motivate automated alternatives. This paper investigates whether Vision-Language Models (VLMs) can assist in assessing perceived image appearance and quality. We introduce a psychophysics-inspired framework to probe VLM perceptual sensitivity through controlled pairwise image comparisons of contrast, colorfulness, and overall preference. Six VLMs (four proprietary and two open-weight models) are compared against psychophysical data. Results reveal strong attribute-dependent variability: Claude exhibits the highest internal consistency, whereas GPT achieves the strongest agreement for overall preference. Claude and Qwen show the strongest alignment for colorfulness, while Qwen performs best for contrast. However, no model consistently matches human perception across all attributes. High self-consistency does not necessarily imply perceptual validity, and VLM--human agreement generally improves when perceptual differences among renderings are more pronounced.

Imran Mehmood, Imad Ali Shah, Ming Ronnier Luo +1

School of Engineering, University of Galway, Ireland · Lero the Research Ireland Centre for Software, Ireland · State Key Laboratory of Extreme Photonics and Instrumentation, Zhejiang University, Hangzhou, China.

Image Quality Assessment VLM Evaluation Vision-Language Alignment

Mar 25, 2026

VFIG: Vectorizing Complex Figures in SVG with Vision-Language Models

Scalable Vector Graphics (SVG) are essential for technical illustration and digital design, offering resolution independence and semantic editability. In practice, original vector files are frequently lost, leaving only rasterized versions (e.g., PNG, JPEG) that resist modification, while manual reconstruction is prohibitively expensive. Progress on automating raster-to-SVG conversion has been bottlenecked by two gaps: existing SVG datasets are dominated by icons and decorative graphics that lack the complexity of professional diagrams, and existing benchmarks rely on pixel- or embedding-level similarity that fails to capture structural correctness (e.g., broken connectivity, misplaced arrows). We close both gaps with paired contributions targeting diagram-centric figures (e.g., model architectures, flowcharts, schematics). For training, we introduce VFIG-Data, the largest figure-to-SVG dataset of its kind at 66K pairs, combining real paper figures converted via a describe-and-generate pipeline with programmatic diagrams that supply noise-free supervision over arrow styles, fonts, and geometry. For evaluation, we introduce VFIG-Bench, a structure-aware evaluation suite, paired with VFIG-Bench-OOD, an out-of-distribution set of figures manually curated from highly cited arXiv papers. Beyond pixel and embedding similarity, our protocol reports rubric-based VLM-Judge scores and Elo ratings from pairwise human preference evaluation. Built on these contributions, VFIG is a VLM family trained with a simple-to-complex SFT curriculum followed by RL with rendering-aware rewards. VFIG achieves state-of-the-art open-source performance, outperforming the best open-source VLM baseline by over 30%, and matches Claude Sonnet 4.6 on VFIG-BENCH: Gemini-Judge 78.2% vs. 76.7% and GPT-Judge 87.5% vs. 87.4%. It remains slightly behind the strongest proprietary models GPT-5.2 and Gemini-3.

Qijia He, Xunmei Liu, Hammaad Memon +6

University of Washington · Allen Institute for Artificial Intelligence · UNC-Chapel Hill

Vector Graphics Generation Vision-Language Models Image-to-Image Translation VLM Evaluation

Mar 24, 2026

Ukrainian Visual Word Sense Disambiguation Benchmark

This study presents a benchmark for evaluating the Visual Word Sense Disambiguation (Visual-WSD) task in Ukrainian. The main goal of the Visual-WSD task is to identify, with minimal contextual information, the most appropriate representation of a given ambiguous word from a set of ten images. To construct this benchmark, we followed a methodology similar to that proposed by (CITATION), who previously introduced benchmarks for the Visual-WSD task in English, Italian, and Farsi. This approach allows us to incorporate the Ukrainian benchmark into a broader framework for cross-language model performance comparisons. We collected the benchmark data semi-automatically and refined it with input from domain experts. We then assessed eight multilingual and multimodal large language models using this benchmark. All tested models performed worse than the zero-shot CLIP-based baseline model (CITATION) used by (CITATION) for the English Visual-WSD task. Our analysis revealed a significant performance gap in the Visual-WSD task between Ukrainian and English.

Yurii Laba, Yaryna Mohytych, Ivanna Rohulia +4

Ukrainian Catholic University 2A Kozelnytska st., Lviv, Ukraine, 79026

VLM Evaluation Multilingual VLM Evaluation

Mar 24, 2026

MedObvious: Exposing the Medical Moravec's Paradox in VLMs via Clinical Triage

Vision Language Models (VLMs) are increasingly used for tasks like medical report generation and visual question answering. However, fluent diagnostic text does not guarantee safe visual understanding. In clinical practice, interpretation begins with pre-diagnostic sanity checks: verifying that the input is valid to read (correct modality and anatomy, plausible viewpoint and orientation, and no obvious integrity violations). Existing benchmarks largely assume this step is solved, and therefore miss a critical failure mode: a model can produce plausible narratives even when the input is inconsistent or invalid. We introduce MedObvious, a 1,880-task benchmark that isolates input validation as a set-level consistency capability over small multi-panel image sets: the model must identify whether any panel violates expected coherence. MedObvious spans five progressive tiers, from basic orientation/modality mismatches to clinically motivated anatomy/viewpoint verification and triage-style cues, and includes five evaluation formats to test robustness across interfaces. Evaluating 17 different VLMs, we find that sanity checking remains unreliable: several models hallucinate anomalies on normal (negative-control) inputs, performance degrades when scaling to larger image sets, and measured accuracy varies substantially between multiple-choice and open-ended settings. These results show that pre-diagnostic verification remains unsolved for medical VLMs and should be treated as a distinct, safety-critical capability before deployment.

Ufaq Khan, Umair Nawaz, L D M S S Teja +5

Mohamed bin Zayed University of Artificial Intelligence, UAE · National Institute of Technology, Silchar · Birmingham City University, UK

VLM Evaluation Medical VLMs VLM Robustness

Mar 24, 2026

ARGENT: Adaptive Hierarchical Image-Text Representations

Large-scale Vision-Language Models (VLMs) such as CLIP learn powerful semantic representations but operate in Euclidean space, which fails to capture the inherent hierarchical structure of visual and linguistic concepts. Hyperbolic geometry, with its exponential volume growth, offers a principled alternative for embedding such hierarchies with low distortion. However, existing hyperbolic VLMs use entailment losses that are unstable: as parent embeddings contract toward the origin, their entailment cones widen toward a half-space, causing catastrophic cone collapse that destroys the intended hierarchy. Additionally, hierarchical evaluation of these models remains unreliable, being largely retrieval-based and correlation-based metrics and prone to taxonomy dependence and ambiguous negatives. To address these limitations, we propose an adaptive entailment loss paired with a norm regularizer that prevents cone collapse without heuristic aperture clipping. We further introduce an angle-based probabilistic entailment protocol (PEP) for evaluating hierarchical understanding, scored with AUC-ROC and Average Precision. This paper introduces a stronger hyperbolic VLM baseline ARGENT, Adaptive hieRarchical imaGe-tExt represeNTation. ARGENT improves the SOTA hyperbolic VLM by 0.7, 1.1, and 0.8 absolute points on image classification, text-to-image retrieval, and proposed hierarchical metrics, respectively.

Chuong Huynh, Hossein Souri, Abhinav Kumar +3

University of Maryland, College Park · Samsung Research America, AI Center – Mountain View

Vision-Language Models Hierarchical Representation Learning Hyperbolic Representation Learning VLM Evaluation

Mar 24, 2026

Quantized Vision-Language Models for Damage Assessment: A Comparative Study of LLaVA-1.5-7B Quantization Levels

Bridge infrastructure inspection is a critical but labor-intensive task requiring expert assessment of structural damage such as rebar exposure, cracking, and corrosion. This paper presents a comprehensive study of quantized Vision-Language Models (VLMs) for automated bridge damage assessment, focusing on the trade-offs between description quality, inference speed, and resource requirements. We develop an end-to-end pipeline combining LLaVA-1.5-7B for visual damage analysis, structured JSON extraction, and rule-based priority scoring. To enable deployment on consumer-grade GPUs, we conduct a systematic comparison of three quantization levels: Q4_K_M, Q5_K_M, and Q8_0 across 254 rebar exposure images. We introduce a 5-point quality evaluation framework assessing damage type recognition, severity classification. Our results demonstrate that Q5_K_M achieves the optimal balance: quality score 3.18

\pm

1.35/5.0, inference time 5.67s/image, and 0.56 quality/sec efficiency -- 8.5% higher quality than Q4_K_M with only 4.5% speed reduction, while matching Q8_0's quality with 25% faster inference. Statistical analysis reveals Q5_K_M exhibits the weakest text-quality correlation (-0.148), indicating consistent performance regardless of description length.

Takato Yasuno

VLM Quantization VLM Evaluation Structural Health Monitoring

Mar 24, 2026

Edge Reliability Gap in Vision-Language Models: Quantifying Failure Modes of Compressed VLMs Under Visual Corruption

The rapid compression of large vision-language models (VLMs) for edge deployment raises an underexplored question: do compact models fail differently, not merely more often? This study compares a 7-billion-parameter quantised VLM (Qwen2.5-VL-7B, 4-bit NF4) against a 500-million-parameter FP16 model (SmolVLM2-500M) across 4,000 samples from VQAv2 and COCO Captions. A three-category error taxonomy (Object Blindness, Semantic Drift, Prior Bias) is applied as a diagnostic framework. A text-only GPT-4o judge reveals Semantic Drift (B) as the dominant failure mode on VQAv2 and on COCO for Qwen, with a mixed Object Blindness / Semantic Drift profile for SmolVLM2 on COCO; Prior Bias (C) is present on VQAv2 but absent on COCO for both models. Confidence calibration is measured via Expected Calibration Error (ECE) using geometric mean token probability, compositional reasoning is probed with structured negation probes across four templates, and a blur robustness experiment completes the evaluation. For this model pair, the compact model exhibits a qualitatively distinct failure signature: a 12.5pp larger negation collapse (-33.2pp vs. -20.8pp, Wald 95% CI [8.2, 16.8]pp, p < 10^-8), driven almost entirely by COCO while the VQAv2 gap is not statistically significant (4.5pp, p=0.19). The most discriminating template is false_yn: SMOLVLM2-500M responds "Yes" (incorrectly claiming a depicted object is absent) on 100% of COCO trials vs. 14% for Q WEN 2.5-VL-7B. Asymmetric dataset-dependent miscalibration and a blur experiment with two controlled ablations complete the analysis. The fully reproducible pipeline is released for systematic safety auditing of compressed VLMs prior to edge deployment.

Mehmet Kaan Erol

Marmara University, Institute of Pure and Applied Sciences

Compositional Reasoning in VLMs VLM Evaluation VLM Robustness VLM Reasoning VLM Quantization VLM Calibration

Mar 24, 2026

TreeTeaming: Autonomous Red-Teaming of Vision-Language Models via Hierarchical Strategy Exploration

The rapid advancement of Vision-Language Models (VLMs) has brought their safety vulnerabilities into sharp focus. However, existing red teaming methods are fundamentally constrained by an inherent linear exploration paradigm, confining them to optimizing within a predefined strategy set and preventing the discovery of novel, diverse exploits. To transcend this limitation, we introduce TreeTeaming, an automated red teaming framework that reframes strategy exploration from static testing to a dynamic, evolutionary discovery process. At its core lies a strategic Orchestrator, powered by a Large Language Model (LLM), which autonomously decides whether to evolve promising attack paths or explore diverse strategic branches, thereby dynamically constructing and expanding a strategy tree. A multimodal actuator is then tasked with executing these complex strategies. In the experiments across 12 prominent VLMs, TreeTeaming achieves state-of-the-art attack success rates on 11 models, outperforming existing methods and reaching up to 87.60% on GPT-4o. The framework also demonstrates superior strategic diversity over the union of previously public jailbreak strategies. Furthermore, the generated attacks exhibit an average toxicity reduction of 23.09%, showcasing their stealth and subtlety. Our work introduces a new paradigm for automated vulnerability discovery, underscoring the necessity of proactive exploration beyond static heuristics to secure frontier AI models.

Chunxiao Li, Lijun Li, Jing Shao

Shanghai Artificial Intelligence Laboratory

Cultural Safety Evaluation of VLMs Automated Red Teaming Adversarial Attacks on VLMs VLM Evaluation

Mar 24, 2026

How Far Can VLMs Go for Visual Bug Detection? Studying 19,738 Keyframes from 41 Hours of Gameplay Videos

Video-based quality assurance (QA) for long-form gameplay video is labor-intensive and error-prone, yet valuable for assessing game stability and visual correctness over extended play sessions. Vision language models (VLMs) promise general-purpose visual reasoning capabilities and thus appear attractive for detecting visual bugs directly from video frames. Recent benchmarks suggest that VLMs can achieve promising results in detecting visual glitches on curated datasets. Building on these findings, we conduct a real-world study using industrial QA gameplay videos to evaluate how well VLMs perform in practical scenarios. Our study samples keyframes from long gameplay videos and asks a VLM whether each keyframe contains a bug. Starting from a single-prompt baseline, the model achieves a precision of 0.50 and an accuracy of 0.72. We then examine two common enhancement strategies used to improve VLM performance without fine-tuning: (1) a secondary judge model that re-evaluates VLM outputs, and (2) metadata-augmented prompting through the retrieval of prior bug reports. Across 100 videos totaling 41 hours and 19,738 keyframes, these strategies provide only marginal improvements over the simple baseline, while introducing additional computational cost and output variance. Our findings indicate that off-the-shelf VLMs are already capable of detecting a certain range of visual bugs in QA gameplay videos, but further progress likely requires hybrid approaches that better separate textual and visual anomaly detection.

Wentao Lu, Alexander Senchenko, Alan Sayle +2

University of Alberta Edmonton, Alberta, Canada · Electronic Arts Vancouver, British Columbia, Canada · Electronic Arts Guildford, Surrey, United Kingdom

VLM Evaluation Video Anomaly Detection

Mar 23, 2026

To Agree or To Be Right? The Grounding-Sycophancy Tradeoff in Medical Vision-Language Models

Vision-language models (VLMs) adapted to the medical domain have shown strong performance on visual question answering benchmarks, yet their robustness against two critical failure modes, hallucination and sycophancy, remains poorly understood, particularly in combination. We evaluate six VLMs (three general-purpose, three medical-specialist) on three medical VQA datasets and uncover a grounding-sycophancy tradeoff: models with the lowest hallucination propensity are the most sycophantic, while the most pressure-resistant model hallucinates more than all medical-specialist models. To characterize this tradeoff, we propose three metrics: L-VASE, a logit-space reformulation of VASE that avoids its double-normalization; CCS, a confidence-calibrated sycophancy score that penalizes high-confidence capitulation; and Clinical Safety Index (CSI), a unified safety index that combines grounding, autonomy, and calibration via a geometric mean. Across 1,151 test cases, no model achieves a CSI above 0.35, indicating that none of the evaluated 7-8B parameter VLMs is simultaneously well-grounded and robust to social pressure. Our findings suggest that joint evaluation of both properties is necessary before these models can be considered for clinical use. Code is available at https://github.com/UTSA-VIRLab/AgreeOrRight

OFM Riaz Rahman Aranya, Kevin Desai

Department of Computer Science, The University of Texas at San Antonio

Sycophancy in VLMs Medical VLMs VLM Evaluation VLM Hallucination VLM Calibration

Mar 23, 2026

SteelDefectX: A Multi-Form Vision-Language Dataset and Benchmark for Steel Surface Defect Analysis

Steel surface defect analysis is critical for industrial quality control, yet existing benchmarks rely primarily on label-only annotations, limiting fine-grained semantic understanding and systematic evaluation of vision-language models. To address this gap, we introduce SteelDefectX, a vision-language dataset with multi-form textual annotations for steel surface defect analysis, comprising 7,778 images across 25 defect categories. At the class level, the dataset provides defect names, representative visual attributes, and industrial causes. At the sample level, each image is annotated with three forms of textual representations: (1) free-form natural language descriptions, (2) structured attribute annotations, and (3) template-based sentences. These annotations provide flexible textual supervision with varying levels of expressiveness and controllability. We further establish a comprehensive benchmark covering vision-language classification, segmentation, and cross-dataset transfer, along with additional evaluations such as retrieval and text-guided localization. Experimental results reveal a trade-off between structure and flexibility in textual representations. Structured attributes provide more stable semantic alignment, while natural language descriptions improve transferability and fine-grained spatial grounding. These findings highlight the critical role of textual design in industrial vision-language learning. SteelDefectX provides a new benchmark for studying semantic alignment and generalization in industrial vision-language learning. The code and dataset are available at https://github.com/Zhaosxian/SteelDefectX.

Shuxian Zhao, Jie Gui, Baosheng Yu +1

Southeast University · Purple Mountain Laboratories · Nanyang Technological University

VLM Evaluation Vision-Language Grounding Industrial Inspection Cross-Dataset Generalization

Mar 23, 2026

MIRAGE: The Illusion of Visual Understanding

Multimodal AI systems have achieved remarkable performance across a broad range of real-world tasks, yet the mechanisms underlying visual-language reasoning remain surprisingly poorly understood. We report three findings that challenge prevailing assumptions about how these systems process and integrate visual information. First, Frontier models readily generate detailed image descriptions and elaborate reasoning traces, including pathology-biased clinical findings, for images never provided; we term this phenomenon mirage reasoning. Second, without any image input, models also attain strikingly high scores across general and medical multimodal benchmarks, bringing into question their utility and design. In the most extreme case, our model achieved the top rank on a standard chest X-ray question-answering benchmark without access to any images. Third, when models were explicitly instructed to guess answers without image access, rather than being implicitly prompted to assume images were present, performance declined markedly. Explicit guessing appears to engage a more conservative response regime, in contrast to the mirage regime in which models behave as though images have been provided. These findings expose fundamental vulnerabilities in how visual-language models reason and are evaluated, pointing to an urgent need for private benchmarks that eliminate textual cues enabling non-visual inference, particularly in medical contexts where miscalibrated AI carries the greatest consequence. We introduce B-Clean as a principled solution for fair, vision-grounded evaluation of multimodal AI systems.

Mohammad Asadi, Jack W. O'Sullivan, Fang Cao +5

Department of Electrical Engineering, Stanford University, CA, USA · Division of Cardiology, Department of Medicine, Stanford University, CA, USA · Department of Biomedical Data Science, Stanford University, CA, USA +3

VLM Evaluation Faithfulness in VLMs Medical VLMs Radiology VLMs

Mar 23, 2026

When Visuals Aren't the Problem: Evaluating Vision-Language Models on Misleading Data Visualizations

Visualizations help communicate data insights, but deceptive data representations can distort their interpretation and propagate misinformation. While recent Vision Language Models (VLMs) perform well on many chart understanding tasks, their ability to detect misleading visualizations, especially when deception arises from subtle reasoning errors in captions, remains poorly understood. Here, we evaluate VLMs on misleading visualization-caption pairs grounded in a fine-grained taxonomy of reasoning errors (e.g., Cherry-picking, Causal inference) and visualization design errors (e.g., Truncated axis, Dual axis, inappropriate encodings). To this end, we develop a benchmark that combines real-world visualization with human-authored, curated misleading captions designed to elicit specific reasoning and visualization error types, enabling controlled analysis across error categories and modalities of misleadingness. Evaluating many commercial and open-source VLMs, we find that models detect visual design errors substantially more reliably than reasoning-based misinformation, and frequently misclassify non-misleading visualizations as deceptive. Overall, our work fills a gap between coarse detection of misleading content and the attribution of the specific reasoning or visualization errors that give rise to it.

Harsh Nishant Lalai, Raj Sanjay Shah, Hanspeter Pfister +2

Birla Institute of Technology and Science, Pilani · Georgia Institute of Technology · Harvard University

VLM Evaluation Chart Understanding

Mar 23, 2026

Rethinking Visual Privacy: A Compositional Privacy Risk Framework for Severity Assessment with VLMs

Existing visual privacy benchmarks largely treat privacy as a binary property, labeling images as private or non-private based on visible sensitive content. We argue that privacy is fundamentally compositional. Attributes that are benign in isolation may combine to produce severe privacy violations. We introduce the Compositional Privacy Risk Taxonomy (CPRT), a regulation-aware framework that organizes visual attributes according to standalone identifiability and compositional harm potential. CPRT defines four graded severity levels and is paired with an interpretable scoring function that assigns continuous privacy severity scores. We further construct a taxonomy-aligned dataset of 6.7K images and derive compositional risk scores. By evaluating frontier and open-weight VLMs we find that frontier models align well with compositional severity when provided structured guidance, but systematically underestimate composition-driven risks. Smaller models struggle to internalize graded privacy reasoning. To bridge this gap, we introduce a deployable 8B SFT model that closely matches frontier-level performance on compositional privacy assessment

Efthymios Tsaprazlis, Tiantian Feng, Anil Ramakrishna +3

University of Southern California, Los Angeles CA 90089, USA · Meta Superintelligence Labs · Amazon AGI

VLM Evaluation Privacy Auditing Compositional Reasoning in VLMs VLM Reasoning Privacy Protection in VLMs

Mar 22, 2026

Many Dialects, Many Languages, One Cultural Lens: Evaluating Multilingual VLMs for Bengali Culture Understanding Across Historically Linked Languages and Regional Dialects

Bangla culture is richly expressed through region, dialect, history, food, politics, media, and everyday visual life, yet it remains underrepresented in multimodal evaluation. To address this gap, we introduce BanglaVerse, a culturally grounded benchmark for evaluating multilingual vision-language models (VLMs) on Bengali culture across historically linked languages and regional dialects. Built from 1,152 manually curated images across nine domains, the benchmark supports visual question answering and captioning, and is expanded into four languages and five Bangla dialects, yielding ~32.2K artifacts. Our experiments show that evaluating only standard Bangla overestimates true model capability: performance drops under dialectal variation, especially for caption generation, while historically linked languages such as Hindi and Urdu retain some cultural meaning but remain weaker for structured reasoning. Across domains, the main bottleneck is missing cultural knowledge rather than visual grounding alone, with knowledge-intensive categories. These findings position BanglaVerse as a more realistic test bed for measuring culturally grounded multimodal understanding under linguistic variation.

Nurul Labib Sayeedi, Md. Faiyaz Abdullah Sayeedi, Shubhashis Roy Dipta +6

VLM Evaluation Multilingual VLM Evaluation Vision-Language Models VLM Robustness

Mar 22, 2026

Consistent but Dangerous: Per-Sample Safety Classification Reveals False Reliability in Medical Vision-Language Models

Consistency under paraphrase, the property that semantically equivalent prompts yield identical predictions, is increasingly used as a proxy for reliability when deploying medical vision-language models (VLMs). We show this proxy is fundamentally flawed: a model can achieve perfect consistency by relying on text patterns rather than the input image. We introduce a four-quadrant per-sample safety taxonomy that jointly evaluates consistency (stable predictions across paraphrased prompts) and image reliance (predictions that change when the image is removed). Samples are classified as Ideal (consistent and image-reliant), Fragile (inconsistent but image-reliant), Dangerous (consistent but not image-reliant), or Worst (inconsistent and not image-reliant). Evaluating five medical VLM configurations across two chest X-ray datasets (MIMIC-CXR, PadChest), we find that LoRA fine-tuning dramatically reduces flip rates but shifts a majority of samples into the Dangerous quadrant: LLaVA-Rad Base achieves a 1.5% flip rate on PadChest while 98.5% of its samples are Dangerous. Critically, Dangerous samples exhibit high accuracy (up to 99.6%) and low entropy, making them invisible to standard confidence-based screening. We observe a negative correlation between flip rate and Dangerous fraction (r = -0.89, n=10) and recommend that deployment evaluations always pair consistency checks with a text-only baseline: a single additional forward pass that exposes the false reliability trap.

Binesh Sadanandan, Vahid Behzadan

SAIL Lab, University of New Haven, West Haven, CT, USA

VLM Evaluation Medical VLMs Faithfulness in VLMs Prompt Sensitivity

Mar 20, 2026

When Negation Is a Geometry Problem in Vision-Language Models

Joint Vision-Language Embedding models such as CLIP typically fail at understanding negation in text queries, for example, failing to distinguish "no" in the query: "a plain blue shirt with no logos". Prior work has largely addressed this limitation through data-centric approaches, fine-tuning CLIP on large-scale synthetic negation datasets. However, these efforts are commonly evaluated using retrieval-based metrics that cannot reliably reflect whether negation is actually understood. In this paper, we identify two key limitations of such evaluation metrics and investigate an alternative evaluation framework based on Multimodal LLMs-as-a-judge, which typically excel at understanding simple yes/no questions about image content, providing a fair evaluation of negation understanding in CLIP models. We then ask whether there already exists a direction in the CLIP embedding space associated with negation. We find evidence that such a direction exists, and show that it can be manipulated through test-time intervention via representation engineering to steer CLIP toward negation-aware behavior without any fine-tuning. Finally, we test negation understanding on non-common image-text samples to evaluate generalization under distribution shifts. Code is at https://github.com/fawazsammani/negation-steering

Fawaz Sammani, Tzoulio Chamiti, Paul Gavrikov +1

ETRO Department, Vrije Universiteit Brussel, Belgium · imec, Kapeldreef 75, B-3001 Leuven, Belgium · Independent Researcher

VLM Reasoning VLM Evaluation VLM Robustness Representation Engineering

Mar 20, 2026

CREG: Compass Relational Evidence Graph for Characterizing Directional Structure in VLM Spatial-Reasoning Attribution

Standard attribution heatmaps show where a vision-language model (VLM) focuses, but they do not reveal whether the recovered evidence is organized by the queried spatial relation or merely reflects image layout. To address this problem, we introduce CREG (Compass Relational Evidence Graph), a training-free diagnostic framework that converts token-level attribution into a reference-centered compass distribution and measures its directional alignment. CREG provides a shared directional readout across attribution methods and makes comparison with geometric controls explicit. Across three spatial-relation benchmarks, box-only geometry achieves Direction Alignment Error 28.4 to 34.4 degrees lower than the best current model-based attribution method on each dataset, leaving a substantial gap between attribution structure and simple target localization. To examine this gap, we apply a diagnostic battery including target intervention, reference-center randomization, and variance partition. Taken together, the results suggest that the directional structure recoverable from current attribution methods is limited and often mixed with image layout. We further find that higher task accuracy does not reliably coincide with better directional attribution: small-scale LoRA training and newer model generations can improve task accuracy while leaving Direction Alignment Error unchanged or worse. These findings characterize what current attribution methods reveal rather than the model's internal spatial representation. CREG provides a controlled protocol for testing whether improvements in spatial reasoning are accompanied by more directionally organized evidence.

Kaizhen Tan, Yang Feng, Heqing Du

Carnegie Mellon University Pittsburgh, PA, USA · Columbia University New York, NY, USA

VLM Evaluation Gradient-Based Attribution VLM Interpretability Spatial Reasoning Spatial Reasoning in Language Models

Mar 20, 2026

From Masks to Pixels and Meaning: A New Taxonomy, Benchmark, and Metrics for VLM Image Tampering

Existing tampering detection benchmarks largely rely on object masks, which severely misalign with the true edit signal: many pixels inside a mask are untouched or only trivially modified, while subtle yet consequential edits outside the mask are treated as natural. We reformulate VLM image tampering from coarse region labels to a pixel-grounded, meaning and language-aware task. First, we introduce a taxonomy spanning edit primitives (replace/remove/splice/inpaint/attribute/colorization, etc.) and their semantic class of tampered object, linking low-level changes to high-level understanding. Second, we release a new benchmark with per-pixel tamper maps and paired category supervision to evaluate detection and classification within a unified protocol. Third, we propose a training framework and evaluation metrics that quantify pixel-level correctness with localization to assess confidence or prediction on true edit intensity, and further measure tamper meaning understanding via semantics-aware classification and natural language descriptions for the predicted regions. We also re-evaluate the existing strong segmentation/localization baselines on recent strong tamper detectors and reveal substantial over- and under-scoring using mask-only metrics, and expose failure modes on micro-edits and off-mask changes. Our framework advances the field from masks to pixels, meanings and language descriptions, establishing a rigorous standard for tamper localization, semantic classification and description. Code and benchmark data are available at https://github.com/VILA-Lab/PIXAR.

Xinyi Shang, Yi Tang, Jiacheng Cui +9

Mohamed bin Zayed University of Artificial Intelligence · University College London

Image Forgery Detection Image Forgery Localization VLM Evaluation

Mar 20, 2026

FREAK: A Fine-grained Hallucination Evaluation Benchmark for Advanced MLLMs

Multimodal Large Language Models (MLLMs) suffer from hallucinations. Existing hallucination evaluation benchmarks are often limited by over-simplified tasks leading to saturated metrics, or insufficient diversity that fails to adequately assess the hallucination extent in state-of-the-art multimodal models. To address this gap, we propose FREAK, a comprehensive multimodal benchmark designed for fine-grained hallucination assessment in MLLMs. Through high-quality photorealistic images featuring fine-grained counter-commonsense edits, FREAK innovatively evaluates hallucination phenomena in detailed visual perception of MLLMs. Extensive experiments on FREAK show severe hallucination issues in SOTA models regarding detailed visual perception. To enable deeper investigation, we curate a controlled subset to indirectly evaluate the model's ability to perceive target detailed information. Through systematic evaluation of prevailing Chain-of-Thought (CoT) prompting techniques within this task, we reveal critical insights regarding hallucination patterns and model reasoning processes.

Zhihan Yin, Jianxin Liang, Yueqian Wang +3

Wangxuan Institute of Computer Technology, Peking University · State Key Laboratory of General Artificial Intelligence

VLM Hallucination VLM Evaluation Counterfactual Evaluation of VLMs

Mar 19, 2026

Do VLMs Need Vision Transformers? Evaluating State Space Models as Vision Encoders

Large vision--language models (VLMs) often use a frozen vision backbone, whose image features are mapped into a large language model through a lightweight connector. While transformer-based encoders are the standard visual backbone, we ask whether state space model (SSM) vision backbones can be a strong alternative. We systematically evaluate SSM vision backbones for VLMs in a controlled setting. Under matched ImageNet-1K initialization, the SSM backbone achieves the strongest overall performance across both VQA and grounding/localization. We further adapt both SSM and ViT-family backbones with detection or segmentation training and find that dense-task tuning generally improves performance across families; after this adaptation, the SSM backbone remains competitive while operating at a substantially smaller model scale. We further observe that (i) higher ImageNet accuracy or larger backbones do not reliably translate into better VLM performance, and (ii) some visual backbones are unstable in localization. Based on these findings, we propose stabilization strategies that improve robustness for both backbone families and highlight SSM backbones as a strong alternative to transformer-based vision encoders in VLMs.

Shang-Jui Ray Kuo, Paola Cascante-Bonilla

Stony Brook University

Vision-Language Models State Space Models VLM Evaluation VLM Robustness

Mar 19, 2026

SAVeS: Steering Safety Judgments in Vision-Language Models via Semantic Cues

Vision-language models (VLMs) are increasingly deployed in real-world and embodied settings where safety decisions depend on visual context. However, it remains unclear which visual evidence drives these judgments. We study whether multimodal safety behavior in VLMs can be steered by simple semantic cues. We introduce a semantic steering framework that applies controlled textual, visual, and cognitive interventions without changing the underlying scene content. To evaluate these effects, we propose SAVeS, a benchmark for situational safety under semantic cues, together with an evaluation protocol that separates behavioral refusal, grounded safety reasoning, and false refusals. Experiments across multiple VLMs and an additional state-of-the-art benchmark show that safety decisions are highly sensitive to semantic cues, indicating reliance on learned visual-linguistic associations rather than grounded visual understanding. We further demonstrate that automated steering pipelines can exploit these mechanisms, highlighting a potential vulnerability in multimodal safety systems.

Carlos Hinojosa, Clemens Grange, Bernard Ghanem

King Abdullah University of Science and Technology (KAUST), Saudi Arabia · Technical University of Munich (TUM), Germany

Cultural Safety Evaluation of VLMs Vision-Language Models VLM Evaluation VLM Robustness

Mar 19, 2026

MultihopSpatial: Multi-hop Compositional Spatial Reasoning Benchmark for Vision-Language Model

Spatial reasoning is foundational for Vision-Language Models (VLMs), particularly when deployed as Vision-Language-Action (VLA) agents in physical environments. However, existing benchmarks predominantly focus on elementary, single-hop relations, neglecting the multi-hop compositional reasoning and precise visual grounding essential for real-world scenarios. To address this, we introduce MultihopSpatial, offering three key contributions: (1) A comprehensive benchmark designed for multi-hop and compositional spatial reasoning, featuring 1- to 3-hop complex queries across diverse spatial perspectives. (2) Acc@50IoU, a complementary metric that simultaneously evaluates reasoning and visual grounding by requiring both answer selection and precise bounding box prediction - capabilities vital for robust VLA deployment. (3) MultihopSpatial-Train, a dedicated large-scale training corpus to foster spatial intelligence. Extensive evaluation of 37 state-of-the-art VLMs yields eight key insights, revealing that compositional spatial reasoning remains a formidable challenge. Finally, we demonstrate that reinforcement learning post-training on our corpus enhances both intrinsic VLM spatial reasoning and downstream embodied manipulation performance.

Youngwan Lee, Soojin Jang, Yoorhim Cho +3

Electronics and Telecommunications Research Institute, South Korea · Korea Advanced Institute of Science and Technology, South Korea · Sungkyunkwan University, South Korea +1

Spatial Reasoning Benchmarks Spatial Reasoning VLM Evaluation VLMs for Robotics Visual Grounding

Mar 19, 2026

Do Vision Language Models Understand Human Engagement in Games?

Inferring human engagement from gameplay video is important for game design and player-experience research, yet it remains unclear whether vision--language models (VLMs) can infer such latent psychological states from visual cues alone. Using the GameVibe Few-Shot dataset across nine first-person shooter games, we evaluate three VLMs under six prompting strategies, including zero-shot prediction, theory-guided prompts grounded in Flow, GameFlow, Self-Determination Theory, and MDA, and retrieval-augmented prompting. We consider both pointwise engagement prediction and pairwise prediction of engagement change between consecutive windows. Results show that zero-shot VLM predictions are generally weak and often fail to outperform simple per-game majority-class baselines. Memory- or retrieval-augmented prompting improves pointwise prediction in some settings, whereas pairwise prediction remains consistently difficult across strategies. Theory-guided prompting alone does not reliably help and can instead reinforce surface-level shortcuts. These findings suggest a perception--understanding gap in current VLMs: although they can recognize visible gameplay cues, they still struggle to robustly infer human engagement across games.

Ziyi Wang, Qizan Guo, Rishitosh Singh +1

Arizona State University

VLM Evaluation Video-Language Models

Mar 17, 2026

Rationale Matters: Learning Transferable Rubrics via Proxy-Guided Critique for VLM Reward Models

Generative reward models (GRMs) for vision-language models (VLMs) often evaluate outputs via a three-stage pipeline: rubric generation, criterion-based scoring, and a final verdict. However, the intermediate rubric is rarely optimized directly. Prior work typically either treats rubrics as incidental or relies on expensive LLM-as-judge checks that provide no differentiable signal and limited training-time guidance. We propose Proxy-GRM, which introduces proxy-guided rubric verification into Reinforcement Learning (RL) to explicitly enhance rubric quality. Concretely, we train lightweight proxy agents (Proxy-SFT and Proxy-RL) that take a candidate rubric together with the original query and preference pair, and then predict the preference ordering using only the rubric as evidence. The proxy's prediction accuracy serves as a rubric-quality reward, incentivizing the model to produce rubrics that are internally consistent and transferable. With ~50k data samples, Proxy-GRM reaches state-of-the-art results on the VL-Reward Bench, Multimodal Reward Bench, and MM-RLHF-Reward Bench, outperforming the methods trained on four times the data. Ablations show Proxy-SFT is a stronger verifier than Proxy-RL, and implicit reward aggregation performs best. Crucially, the learned rubrics transfer to unseen evaluators, improving reward accuracy at test time without additional training. Our code is available at https://github.com/Qwen-Applications/Proxy-GRM.

Weijie Qiu, Dai Guan, Junxin Wang +6

Qwen Large Model Application Team, Alibaba · Institute of Automation, Chinese Academy of Sciences

Multimodal Reward Modeling VLM Evaluation Rubric-Based RL Rubric-Guided Generation RL for VLMs