Pith. sign in

REVIEW 1 major objections 3 minor 60 cited by

Large-scale pre-training of a vision foundation model followed by reinforcement learning with curriculum sampling produces GLM-4.5V, which leads open-source models on nearly all of 42 multimodal benchmarks.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Download PDF Read on arXiv ↗

T0 review · grok-4.3

2026-05-11 04:43 UTC

load-bearing objection GLM-4.5V shows that curriculum RL on a strong base VLM can lift open models to competitive levels on reasoning tasks, with the 9B variant and code released for direct use. the 1 major comments →

arxiv 2507.01006 v6 submitted 2025-07-01 cs.CV cs.AIcs.LG

GLM-4.5V and GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning

GLM-V Team: Wenyi Hong , Wenmeng Yu , Xiaotao Gu , Guo Wang , Guobing Gan , Haomiao Tang , Jiale Cheng , Ji Qi
show 84 more authors
Junhui Ji Lihang Pan Shuaiqi Duan Weihan Wang Yan Wang Yean Cheng Zehai He Zhe Su Zhen Yang Ziyang Pan Aohan Zeng Baoxu Wang Bin Chen Boyan Shi Changyu Pang Chenhui Zhang Da Yin Fan Yang Guoqing Chen Haochen Li Jiale Zhu Jiali Chen Jiaxing Xu Jiazheng Xu Jing Chen Jinghao Lin Jinhao Chen Jinjiang Wang Junjie Chen Leqi Lei Letian Gong Leyi Pan Mingdao Liu Mingde Xu Mingzhi Zhang Qinkai Zheng Ruiliang Lyu Shangqin Tu Sheng Yang Shengbiao Meng Shi Zhong Shiyu Huang Shuyuan Zhao Siyan Xue Tianshu Zhang Tianwei Luo Tianxiang Hao Tianyu Tong Wei Jia Wenkai Li Xiao Liu Xiaohan Zhang Xin Lyu Xinyu Zhang Xinyue Fan Xuancheng Huang Yadong Xue Yanfeng Wang Yanling Wang Yanzi Wang Yifan An Yifan Du Yiheng Huang Yilin Niu Yiming Shi Yu Wang Yuan Wang Yuanchang Yue Yuchen Li Yusen Liu Yutao Zhang Yuting Wang Yuxuan Zhang Zhao Xue Zhengxiao Du Zhenyu Hou Zihan Wang Peng Zhang Debing Liu Bin Xu Juanzi Li Minlie Huang Yuxiao Dong Jie Tang
This is my paper
classification cs.CV cs.AIcs.LG
keywords vision-language modelsmultimodal reasoningreinforcement learningcurriculum samplingGLM-4.5Vopen-source modelsbenchmark evaluationGUI agents
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper first builds a strong vision foundation model through extensive pre-training, which it treats as setting the performance ceiling. It then applies Reinforcement Learning with Curriculum Sampling (RLCS) to progressively train the model on harder examples drawn from a structured curriculum. This combination yields broad gains across STEM problem-solving, video understanding, coding, grounding, GUI agents, and long-document tasks. GLM-4.5V reaches state-of-the-art results among open-source models of comparable size and matches or exceeds closed-source systems such as Gemini-2.5-Flash on coding and agent benchmarks. The smaller GLM-4.1V-9B-Thinking variant even outperforms the much larger Qwen2.5-VL-72B on 29 of the evaluated benchmarks.

Core claim

The central claim is that a capable vision foundation model pre-trained at large scale can have its full potential realized through Reinforcement Learning with Curriculum Sampling, producing versatile multimodal reasoning that improves performance across a wide range of tasks without evident overfitting to specific benchmarks.

What carries the argument

Reinforcement Learning with Curriculum Sampling (RLCS), which samples training examples from a progressively harder curriculum to refine the pre-trained vision-language model for reasoning.

Load-bearing premise

That the large-scale pre-trained vision foundation model already encodes a reliable upper bound on capability and that RLCS can unlock this bound without creating benchmark-specific overfitting or evaluation artifacts.

What would settle it

A new multimodal reasoning benchmark drawn from entirely unseen distributions would show whether GLM-4.5V maintains its reported performance edge or drops to levels comparable with prior open models.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • GLM-4.5V sets new open-source records on nearly all of 42 public benchmarks spanning STEM, video, coding, GUI agents, and document understanding.
  • The 9B GLM-4.1V-Thinking variant surpasses the 72B Qwen2.5-VL on 29 benchmarks despite its smaller size.
  • The models demonstrate competitive or superior results to closed-source Gemini-2.5-Flash specifically on coding and GUI-agent tasks.
  • The GLM-4.6V series adds native tool use and a 128K context window while retaining the same training approach.
  • Open-sourcing the 9B Thinking model and GLM-4.5V enables direct community inspection and further fine-tuning.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Curriculum sampling during RL may prove more important than raw scale for avoiding overfitting in vision-language models.
  • The same pre-train-then-RLCS recipe could be tested on non-vision modalities to check whether the performance pattern generalizes.
  • If the method scales cleanly, future open models might routinely match or exceed closed models on agent-style tasks without requiring proprietary data.
  • The approach highlights a practical way to convert raw pre-training compute into measurable gains on long-horizon reasoning benchmarks.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

1 major / 3 minor

Summary. The paper introduces GLM-4.1V-Thinking, GLM-4.5V, and GLM-4.6V, a family of vision-language models built from a large-scale pre-trained vision foundation model that is further improved via Reinforcement Learning with Curriculum Sampling (RLCS). It reports that GLM-4.5V achieves state-of-the-art results among open-source models of comparable size across 42 public benchmarks and is competitive with or superior to closed-source models such as Gemini-2.5-Flash on coding and GUI-agent tasks; the smaller GLM-4.1V-9B-Thinking outperforms the much larger Qwen2.5-VL-72B on 29 benchmarks. The work supplies training details, ablations, and benchmark tables, and open-sources the GLM-4.1V-9B-Thinking and GLM-4.5V checkpoints together with code.

Significance. If the empirical results hold, the paper demonstrates that scalable RL with curriculum sampling can substantially unlock multimodal reasoning potential in a strong vision foundation model, yielding open-source VLMs that rival or exceed larger open models and some closed systems on diverse tasks. The release of models, code, and detailed training information provides a valuable, reproducible baseline for the community.

major comments (1)
  1. [Section 3] Section 3 (RLCS): the curriculum sampling schedule is described at a high level with free parameters listed in the method; the paper should report sensitivity analysis or default values used for the schedule, as these directly affect reproducibility of the claimed performance gains.
minor comments (3)
  1. [Table 1] Table 1 and benchmark tables: include error bars or standard deviations from multiple runs where available, and explicitly state the data splits or contamination checks performed for the 42 benchmarks.
  2. [Abstract] Abstract and Section 1: the brief mention of GLM-4.6V (native tool use, 128K context) should be expanded with one sentence on how it differs from the 4.5V series to clarify the overall model family.
  3. [Figures] Figure captions and training curves: ensure all axes are labeled with units and that the curves are referenced in the text when discussing ablation results.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the positive evaluation and the recommendation of minor revision. We address the single major comment below.

read point-by-point responses

  1. Referee: [Section 3] Section 3 (RLCS): the curriculum sampling schedule is described at a high level with free parameters listed in the method; the paper should report sensitivity analysis or default values used for the schedule, as these directly affect reproducibility of the claimed performance gains.

    Authors: We agree that additional details on the curriculum sampling schedule parameters would improve reproducibility. In the revised manuscript we will explicitly list the default values employed for all free parameters in the RLCS formulation and include a concise sensitivity analysis on the most impactful hyperparameters, drawing from the ablation experiments already conducted during development. revision: yes

Circularity Check

0 steps flagged

No significant circularity; empirical claims only

full rationale

The paper presents an empirical pipeline: large-scale pre-training of a vision foundation model followed by RLCS training, with performance evaluated on 42 public benchmarks. No mathematical derivations, equations, fitted parameters renamed as predictions, or load-bearing self-citations appear in the provided text. The central claims rest on benchmark results and open-sourced models rather than reducing to inputs by construction. This is the standard non-circular outcome for an applied ML report.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claims rest on standard assumptions of large-scale pretraining and RL optimization; no new physical entities or ad-hoc constants are introduced beyond typical training hyperparameters.

free parameters (1)
  • Curriculum sampling schedule parameters
    The RLCS method requires choices for how task difficulty increases over training; these are not quantified in the abstract but are necessary for the reported gains.
axioms (2)
  • domain assumption Large-scale pre-training produces a vision foundation model whose capabilities form an upper bound for subsequent RL fine-tuning.
    Stated directly in the abstract as the first step before applying RLCS.
  • domain assumption Public multimodal benchmarks provide an unbiased measure of general reasoning capability.
    The evaluation across 42 benchmarks is presented as comprehensive evidence of capability enhancement.

pith-pipeline@v0.9.0 · 5999 in / 1492 out tokens · 65273 ms · 2026-05-11T04:43:49.425790+00:00 · methodology

0 comments
read the original abstract

We present GLM-4.1V-Thinking, GLM-4.5V, and GLM-4.6V, a family of vision-language models (VLMs) designed to advance general-purpose multimodal understanding and reasoning. In this report, we share our key findings in the development of the reasoning-centric training framework. We first develop a capable vision foundation model with significant potential through large-scale pre-training, which arguably sets the upper bound for the final performance. We then propose Reinforcement Learning with Curriculum Sampling (RLCS) to unlock the full potential of the model, leading to comprehensive capability enhancement across a diverse range of tasks, including STEM problem solving, video understanding, content recognition, coding, grounding, GUI-based agents, and long document interpretation. In a comprehensive evaluation across 42 public benchmarks, GLM-4.5V achieves state-of-the-art performance on nearly all tasks among open-source models of similar size, and demonstrates competitive or even superior results compared to closed-source models such as Gemini-2.5-Flash on challenging tasks including Coding and GUI Agents. Meanwhile, the smaller GLM-4.1V-9B-Thinking remains highly competitive-achieving superior results to the much larger Qwen2.5-VL-72B on 29 benchmarks. We open-source both GLM-4.1V-9B-Thinking and GLM-4.5V. We further introduce the GLM-4.6V series, open-source multimodal models with native tool use and a 128K context window. A brief overview is available at https://z.ai/blog/glm-4.6v. Code, models and more information are released at https://github.com/zai-org/GLM-V.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 60 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. DataComp-VLM: Improved Open Datasets for Vision-Language Models

    cs.CV 2026-06 conditional novelty 8.0

    DataComp-VLM benchmark shows instruction-heavy data mixing outperforms filtering for VLM training, with DCVLM-Baseline achieving 63.6% on 33 tasks for 8B models (+5.4pp over FineVision).

  2. MMRareBench: A Rare-Disease Multimodal and Multi-Image Medical Benchmark

    cs.CV 2026-04 unverdicted novelty 8.0

    MMRareBench is the first rare-disease benchmark for multimodal and multi-image clinical evaluation of MLLMs, revealing fragmented capabilities, low treatment-planning scores, and medical models underperforming general...

  3. MMRareBench: A Rare-Disease Multimodal and Multi-Image Medical Benchmark

    cs.CV 2026-04 unverdicted novelty 8.0

    MMRareBench provides 1,756 QA pairs and 7,958 images from PMC rare-disease cases to evaluate 23 MLLMs, revealing low treatment-planning scores and medical models underperforming general models on multi-image tasks due...

  4. HM-Bench: A Comprehensive Benchmark for Multimodal Large Language Models in Hyperspectral Remote Sensing

    cs.CV 2026-04 accept novelty 8.0

    HM-Bench is the first benchmark for MLLMs on hyperspectral images, showing models struggle with complex spatial-spectral reasoning and perform better with visual PCA images than textual reports.

  5. Supply-Chain Poisoning Attacks Against LLM Coding Agent Skill Ecosystems

    cs.CR 2026-04 unverdicted novelty 8.0

    DDIPE poisons LLM agent skills by embedding malicious logic in documentation examples, achieving 11.6-33.5% bypass rates across frameworks while explicit attacks are blocked, with 2.5% evading detection.

  6. Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding

    cs.CV 2026-01 unverdicted novelty 8.0

    Molmo2 delivers state-of-the-art open-weight video VLMs with new grounding datasets and training methods that outperform prior open models and match or exceed some proprietary ones on pointing and tracking tasks.

  7. OmniCoT: A Benchmark for Global and Multi-Step Panoramic Reasoning

    cs.CV 2026-06 unverdicted novelty 7.0

    OmniCoT is a new panoramic reasoning benchmark with 6.7K eval, 1K real, and 14.3K training examples plus a two-stage SFT+GRPO training method to enforce global 360-degree consistency.

  8. MuseBench: Benchmarking Intent-Level Audiovisual Arts Understanding in MLLMs

    cs.CV 2026-06 unverdicted novelty 7.0

    MuseBench shows state-of-the-art MLLMs achieve only 48.29% accuracy on intent-level audiovisual arts understanding versus 87.18% for human experts.

  9. SpreadsheetBench 2: Evaluating Agents on End-to-End Business Spreadsheet Workflows

    cs.SE 2026-06 unverdicted novelty 7.0

    SpreadsheetBench 2 provides 321 expert-validated tasks from authentic business data showing frontier LLMs reach only 34.89% overall accuracy on end-to-end spreadsheet workflows.

  10. Position Rebinding Cache Reuse: Replay-Free Visual Revisiting for Interleaved Multimodal Reasoning

    cs.CV 2026-06 unverdicted novelty 7.0

    PRCR enables replay-free visual revisiting in interleaved multimodal reasoning by storing raw visual KV caches with spatial coordinates and rebinding keys to position-compatible coordinates, matching replay performanc...

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

    cs.CV 2026-06 unverdicted novelty 7.0

    SSMNBench shows that MLLMs suffer distraction degradation on single-view-sufficient tasks and fail to integrate geometric evidence across views, instead relying on semantic averaging and view preference.

  12. HG-Bench: A Benchmark for Multi-Page Handwritten Answer-Region Grounding in Automated Homework Assessment

    cs.CV 2026-06 unverdicted novelty 7.0

    HG-Bench supplies 500 human-annotated homework samples and a page-aware protocol that measures complete-answer localization (FA) and step-level decomposition (FSm), exposing that no zero-shot VLM exceeds 55% on either metric.

  13. Each Judge Its Own Yardstick: Discovering Per-VLM Taxonomies for Physical Video Evaluation

    cs.CV 2026-06 unverdicted novelty 7.0

    JudgeFit produces per-VLM physical video evaluation taxonomies that improve held-out accuracy by a mean 32% relative to a single global schema across 16 models from eight families.

  14. CheXpercept: A Benchmark for Evaluating Expert-Level Lesion Perception in Chest X-rays

    cs.CV 2026-06 unverdicted novelty 7.0

    CheXpercept is a sequential multi-level perception benchmark showing VLMs perform adequately only on coarse lesion detection in chest X-rays while degrading sharply on finer tasks, with medical VLMs offering no advant...

  15. PorTEXTO: A European Portuguese Benchmark for Visual Text Extraction

    cs.CV 2026-06 unverdicted novelty 7.0

    PorTEXTO benchmark shows sharp real-world performance drops in pt-PT OCR and finds specialized multilingual data outperforms model size or resolution increases.

  16. One Polluted Page Is Enough: Evaluating Web Content Pollution in Generative Recommenders

    cs.CL 2026-06 unverdicted novelty 7.0

    FORGE benchmark shows search-augmented LLMs recommend fake products at rates up to 27% from one polluted page and 73.8% from top-3 replacement across 12 models and 225 products.

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

    cs.AI 2026-06 unverdicted novelty 7.0

    SpatialWorld is a new multi-simulator benchmark showing top multimodal agents achieve under 18% success on interactive spatial tasks requiring active exploration and long-horizon planning.

  18. NutriMLLM: Multimodal Large Language Models for Dietary Micronutrient Analysis

    cs.CV 2026-06 unverdicted novelty 7.0

    NutriMLLM models fine-tuned on 1.1 million synthetic food image-nutrient triplets from population dietary recalls achieve near-complete coverage and competitive accuracy on real food images for comprehensive micronutr...

  19. UltraVR: A Diagnostic Ultra-Resolution Image-VQA Benchmark for Evidence-Grounded Reasoning

    cs.CV 2026-06 unverdicted novelty 7.0

    UltraVR is a new diagnostic benchmark for evidence-grounded VQA on ultra-resolution images, with structured chain-of-thought annotations that localize failures in grounding, perception, and inference.

  20. FindIt: A Format-Informed Visual Detection Benchmark for Generalist Multimodal LLMs

    cs.CV 2026-06 unverdicted novelty 7.0

    FindIt is the first comprehensive benchmark for evaluating generalist MLLMs on promptable object detection, referring expression detection, instance-level detection, and video detection with standardized parsable outputs.

  21. Benchmarking Visual State Tracking in Multimodal Video Understanding

    cs.CV 2026-06 unverdicted novelty 7.0

    VSTAT benchmark shows state-of-the-art MLLMs perform far below humans and only modestly above answer-prior baselines on visual state tracking, failing at visual perception despite correct textual reasoning.

  22. Towards Characterizing Scientific Image Utility and Upgradability

    cs.CV 2026-06 unverdicted novelty 7.0

    The SIU²A framework evaluates scientific images for error detection, repair feasibility, and correction quality, showing current multimodal systems have major limitations in preserving scientific validity.

  23. Moment-Video: Diagnosing Temporal Fidelity of Video MLLMs on Momentary Visual Events

    cs.CV 2026-06 conditional novelty 7.0

    Moment-Video benchmark shows top video MLLM achieves only 39.6% accuracy on momentary visual event tasks, with most open-source models below 25%.

  24. ChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats

    cs.CV 2026-05 unverdicted novelty 7.0

    ChartArena is a new benchmark dataset and evaluation protocol for chart parsing by MLLMs that covers numeric and diagrammatic charts in multiple languages and real-world visual conditions.

  25. MM-Snowball: Evaluating and Mitigating Hallucination Snowballing in Multimodal Multi-Turn Dialogue

    cs.CV 2026-05 unverdicted novelty 7.0

    MM-Snowball benchmark diagnoses hallucination snowballing in multi-turn MLLM dialogues; CAVR mitigates it via dual visual rectification at representation and logit levels.

  26. DeepLatent: Think with Images via Parallel Latent Visual Reasoning

    cs.CV 2026-05 unverdicted novelty 7.0

    DeepLatent introduces a parallel latent visual reasoning framework with learnable 2D tokens and continuous RL, trained via distillation then RL, plus a new 180K dataset, claiming SOTA benchmark results.

  27. StemBind: When MLLMs Get Lost Between Rules and Instances in Abstract Visual Reasoning

    cs.CV 2026-05 unverdicted novelty 7.0

    StemBind benchmark diagnoses MLLM failures in abstract visual reasoning by separating perception, rule induction, and answer selection on shared stems, finding a persistent rule-to-instance binding gap even when perce...

  28. Every Act Has Its Price: Compressed Moral Composition in Frontier LLMs

    cs.CL 2026-05 unverdicted novelty 7.0

    Moral Trolley Arena shows frontier LLMs produce composite moral preferences that are compressed rather than additive functions of calibrated component act strengths across Moral Foundations Theory.

  29. OmniMatBench: A Human-Calibrated Multimodal Reasoning Benchmark Across 19 Materials Science Subfields

    cs.AI 2026-05 unverdicted novelty 7.0

    OmniMatBench is a new human-calibrated benchmark for multimodal materials-science reasoning that reveals the best evaluated MLLM scores only 0.372 overall.

  30. POINav: Benchmarking and Enhancing Final-Meters Arrival in Real-World Vision-Language Navigation

    cs.RO 2026-05 unverdicted novelty 7.0

    POINav-Bench provides the first high-fidelity real-world benchmark for POI-goal VLN using 3DGS reconstructions of 126k m² with 163 POIs, supported by a Brain-Action framework and 70K real signage-entrance dataset.

  31. AndroidDaily: A Verifiable Benchmark for Mobile GUI Agents on Real-World Closed-Source Applications

    cs.CV 2026-05 unverdicted novelty 7.0

    AndroidDaily supplies 350 verifiable tasks on 94 closed-source Android apps evaluated by GRADE (87.37% human agreement), with the strongest model achieving 62% success.

  32. Towards Open-World Referring Expression Comprehension: A Benchmark with Training-free Multi-task Consistency Checker

    cs.CV 2026-05 unverdicted novelty 7.0

    OpenRef benchmark for open-world REC with F1 and N3R metrics and training-free MCC to improve existing models in complex scenarios.

  33. AgroTools: A Benchmark for Tool-Augmented Multimodal Agents in Agriculture

    cs.CV 2026-05 unverdicted novelty 7.0

    AgroTools is a new benchmark for tool-augmented multimodal agents in agriculture featuring 539 QA pairs, 1,097 images, five task families, and 14 tools, with evaluations showing major limitations in current models' to...

  34. Perception or Prejudice: Can MLLMs Go Beyond First Impressions of Personality?

    cs.AI 2026-05 unverdicted novelty 7.0

    Introduces the Grounded Personality Reasoning task and MM-OCEAN dataset to show that MLLMs frequently produce correct Big Five personality ratings without grounding them in observable video evidence.

  35. JMed48k: A Multi-Profession Japanese Medical Licensing Benchmark for Vision-Language Model Evaluation

    cs.CV 2026-05 conditional novelty 7.0

    JMed48k is a new large-scale benchmark of Japanese medical licensing exams with images that reveals proprietary VLMs benefit more from visuals than medical-specific models, with large variation across professions.

  36. JMed48k: A Multi-Profession Japanese Medical Licensing Benchmark for Vision-Language Model Evaluation

    cs.CV 2026-05 unverdicted novelty 7.0

    JMed48k is a new benchmark of Japanese healthcare licensing exams used to evaluate 21 VLMs, with a paired image-removal audit revealing large differences in how models and professions benefit from visual content.

  37. ECUAS$_n$: A family of metrics for principled evaluation of uncertainty-augmented systems

    cs.AI 2026-05 unverdicted novelty 7.0

    Introduces the ECUAS_n family of proper scoring rules for evaluating uncertainty-augmented systems, where n tunes the trade-off between prediction accuracy costs and uncertainty quality.

  38. HEED: Density-Weighted Residual Alignment for Hybrid Vision-Language Model Distillation

    cs.CV 2026-05 unverdicted novelty 7.0

    HEED replaces uniform residual alignment with density-weighted alignment using patch self-dissimilarity to improve hybrid VLM distillation, gaining 8.7 points on OCRBench v2 and 5.13 on a 10-benchmark average.

  39. PAGER: Bridging the Semantic-Execution Gap in Point-Precise Geometric GUI Control

    cs.AI 2026-05 unverdicted novelty 7.0

    PAGER achieves 4.1x higher task success in point-precise geometric GUI control by combining topology-aware planning with precision-aligned reinforcement learning on the new PAGE Bench dataset of 4,906 problems.

  40. MemLens: Benchmarking Multimodal Long-Term Memory in Large Vision-Language Models

    cs.CV 2026-05 unverdicted novelty 7.0

    MemLens benchmark shows long-context LVLMs lose accuracy with length while memory agents lose visual fidelity, with multi-session reasoning below 30% for most systems and neither approach solving the task alone.

  41. Known By Their Actions: Fingerprinting LLM Browser Agents via UI Traces

    cs.CR 2026-05 unverdicted novelty 7.0

    UI traces of actions and timings from LLM browser agents enable identification of the underlying model with up to 96% F1 across 14 models and multiple tasks.

  42. Chronicles-OCR: A Cross-Temporal Perception Benchmark for the Evolutionary Trajectory of Chinese Characters

    cs.CV 2026-05 unverdicted novelty 7.0

    Chronicles-OCR is the first benchmark with 2,800 images across the complete evolutionary trajectory of Chinese characters, defining four tasks to evaluate VLLMs' cross-temporal visual perception.

  43. Mem-W: Latent Memory-Native GUI Agents

    cs.CL 2026-05 unverdicted novelty 7.0

    Mem-W embeds historical trajectories and working memory as compact latent tokens into GUI agents' continuous context via a trajectory-to-latent compressor, yielding up to +30 point gains on navigation benchmarks.

  44. UniShield: Unified Face Attack Detection via KG-Informed Multimodal Reasoning

    cs.CV 2026-05 unverdicted novelty 7.0

    UniShield introduces a knowledge-graph-informed multimodal framework that improves unified detection of physical and digital face attacks through instruction tuning and consistency-optimized reasoning.

  45. SalesSim: Benchmarking and Aligning Multimodal Language Models as Retail User Simulators

    cs.CL 2026-05 unverdicted novelty 7.0

    SalesSim benchmarks MLLMs as retail user simulators, finds gaps in persona adherence and over-persuasion, and introduces UserGRPO RL to raise decision alignment by 13.8%.

  46. SphereVAD: Training-Free Video Anomaly Detection via Geodesic Inference on the Unit Hypersphere

    cs.CV 2026-05 unverdicted novelty 7.0

    SphereVAD performs training-free video anomaly detection by recasting anomaly discrimination as von Mises-Fisher likelihood-ratio geodesic inference on the unit hypersphere using intermediate MLLM features, with Frech...

  47. RobotEQ: Transitioning from Passive Intelligence to Active Intelligence in Embodied AI

    cs.RO 2026-05 unverdicted novelty 7.0

    RobotEQ is a new benchmark dataset and evaluation suite showing that current embodied AI models fall short on active social-norm compliance, especially spatial grounding, though RAG with external knowledge helps.

  48. RobotEQ: Transitioning from Passive Intelligence to Active Intelligence in Embodied AI

    cs.RO 2026-05 unverdicted novelty 7.0

    RobotEQ is the first benchmark for active intelligence in embodied AI, demonstrating that current models underperform on social norm adherence and spatial grounding tasks.

  49. MolRecBench-Wild: A Real-World Benchmark for Optical Chemical Structure Recognition

    cs.AI 2026-05 unverdicted novelty 7.0

    MolRecBench-Wild reveals that 18 existing OCSR models suffer severe performance drops on complex real-world academic molecular images compared with prior patent benchmarks.

  50. VT-Bench: A Unified Benchmark for Visual-Tabular Multi-Modal Learning

    cs.CV 2026-05 unverdicted novelty 7.0

    VT-Bench is the first unified benchmark aggregating 14 visual-tabular datasets with over 756K samples and evaluating 23 models to expose challenges in this multi-modal area.

  51. VT-Bench: A Unified Benchmark for Visual-Tabular Multi-Modal Learning

    cs.CV 2026-05 unverdicted novelty 7.0

    VT-Bench aggregates 14 datasets from 9 domains and evaluates 23 models to standardize visual-tabular discriminative and generative tasks.

  52. VT-Bench: A Unified Benchmark for Visual-Tabular Multi-Modal Learning

    cs.CV 2026-05 unverdicted novelty 7.0

    VT-Bench aggregates 14 datasets totaling over 756K samples across 9 domains and evaluates 23 models to establish a unified testbed for visual-tabular multi-modal discriminative and generative tasks.

  53. Purifying Multimodal Retrieval: Fragment-Level Evidence Selection for RAG

    cs.IR 2026-04 unverdicted novelty 7.0

    FES-RAG reframes multimodal RAG as fragment-level selection using Fragment Information Gain to outperform document-level methods with up to 27% relative CIDEr gains on M2RAG while shortening context.

  54. OS-SPEAR: A Toolkit for the Safety, Performance,Efficiency, and Robustness Analysis of OS Agents

    cs.CL 2026-04 unverdicted novelty 7.0

    OS-SPEAR is a new evaluation toolkit that tests 22 OS agents and identifies trade-offs between efficiency and safety or robustness.

  55. Towards Temporal Compositional Reasoning in Long-Form Sports Videos

    cs.CV 2026-04 unverdicted novelty 7.0

    SportsTime benchmark and CoTR method improve multimodal AI's temporal compositional reasoning and evidence grounding in long-form sports videos.

  56. X-PCR: A Benchmark for Cross-modality Progressive Clinical Reasoning in Ophthalmic Diagnosis

    cs.CV 2026-04 unverdicted novelty 7.0

    X-PCR is a new benchmark of 26,415 images and 177,868 expert VQA pairs that evaluates MLLMs on six-stage progressive reasoning and cross-modality integration in ophthalmology.

  57. HyLaR: Hybrid Latent Reasoning with Decoupled Policy Optimization

    cs.CV 2026-04 unverdicted novelty 7.0

    HyLaR with DePO enables effective RL in hybrid discrete-continuous spaces for multimodal models, outperforming prior MLLMs on perception and understanding benchmarks.

  58. OASIS: On-Demand Hierarchical Event Memory for Streaming Video Reasoning

    cs.CV 2026-04 unverdicted novelty 7.0

    OASIS organizes streaming video into hierarchical events and retrieves memory on-demand via intent-driven refinement to improve long-horizon accuracy and compositional reasoning with bounded token costs.

  59. MirrorBench: Evaluating Self-centric Intelligence in MLLMs by Introducing a Mirror

    cs.AI 2026-04 unverdicted novelty 7.0

    MirrorBench reveals that leading MLLMs perform far below humans on tasks requiring self-referential perception and representation, even at the simplest level.

  60. RiskWebWorld: A Realistic Interactive Benchmark for GUI Agents in E-commerce Risk Management

    cs.AI 2026-04 unverdicted novelty 7.0

    RiskWebWorld is the first realistic interactive benchmark for GUI agents in e-commerce risk management, revealing a large gap between generalist and specialized models plus RL gains.

Reference graph

Works this paper leans on

77 extracted references · 77 canonical work pages · cited by 149 Pith papers · 29 internal anchors

  1. [1]

    Flame-code-vlm.https://github.com/Flame-Code-VLM/Flame-Code-VLM

    work page

  2. [2]

    Geobench.https://github.com/ccmdi/geobench

    work page

  3. [3]

    Awadalla, L

    A. Awadalla, L. Xue, O. Lo, M. Shu, H. Lee, E. Guha, S. Shen, M. Awadalla, S. Savarese, C. Xiong, et al. Mint-1t: Scaling open-source multimodal data by 10x: A multimodal dataset with one trillion tokens.Advances in Neural Information Processing Systems, 37:36805–36828, 2024

    work page 2024

  4. [4]

    S. Bai, K. Chen, X. Liu, J. Wang, W. Ge, S. Song, K. Dang, P. Wang, S. Wang, J. Tang, et al. Qwen2. 5-vl technical report.arXiv preprint arXiv:2502.13923, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  5. [5]

    Nougat: Neural Optical Understanding for Academic Documents

    L. Blecher, G. Cucurull, T. Scialom, and R. Stojnic. Nougat: Neural optical understanding for academic documents.arXiv preprint arXiv:2308.13418, 2023

    work page internal anchor Pith review arXiv 2023

  6. [6]

    J. Chen, F. Wei, J. Zhao, S. Song, B. Wu, Z. Peng, S.-H. G. Chan, and H. Zhang. Revisiting referring expression comprehension evaluation in the era of large multimodal models. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 513–524, 2025

    work page 2025

  7. [7]

    L. Chen, J. Li, X. Dong, P. Zhang, Y . Zang, Z. Chen, H. Duan, J. Wang, Y . Qiao, D. Lin, et al. Are we on the right way for evaluating large vision-language models?arXiv preprint arXiv:2403.20330, 2024

    work page internal anchor Pith review arXiv 2024

  8. [8]

    A. Fang, A. M. Jose, A. Jain, L. Schmidt, A. Toshev, and V . Shankar. Data filtering networks. arXiv preprint arXiv:2309.17425, 2023

    work page arXiv 2023

  9. [9]

    E. Fini, M. Shukor, X. Li, P. Dufter, M. Klein, D. Haldimann, S. Aitharaju, V . G. T. da Costa, L. Béthune, Z. Gan, et al. Multimodal autoregressive pre-training of large vision encoders. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 9641–9654, 2025

    work page 2025

  10. [10]

    C. Fu, Y . Dai, Y . Luo, L. Li, S. Ren, R. Zhang, Z. Wang, C. Zhou, Y . Shen, M. Zhang, et al. Video-mme: The first-ever comprehensive evaluation benchmark of multi-modal llms in video analysis.arXiv:2405.21075, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  11. [11]

    X. Fu, Y . Hu, B. Li, Y . Feng, H. Wang, X. Lin, D. Roth, N. A. Smith, W.-C. Ma, and R. Kr- ishna. Blink: Multimodal large language models can see but not perceive.arXiv preprint arXiv:2404.12390, 2024

    work page internal anchor Pith review arXiv 2024

  12. [12]

    S. Y . Gadre, G. Ilharco, A. Fang, J. Hayase, G. Smyrnis, T. Nguyen, R. Marten, M. Wortsman, D. Ghosh, J. Zhang, et al. Datacomp: In search of the next generation of multimodal datasets. Advances in Neural Information Processing Systems, 36:27092–27112, 2023

    work page 2023

  13. [13]

    T. GLM, A. Zeng, B. Xu, B. Wang, C. Zhang, D. Yin, D. Rojas, G. Feng, H. Zhao, H. Lai, et al. Chatglm: A family of large language models from glm-130b to glm-4 all tools.arXiv preprint arXiv:2406.12793, 2024

    work page internal anchor Pith review arXiv 2024

  14. [14]

    J. Gu, X. Meng, G. Lu, L. Hou, N. Minzhe, X. Liang, L. Yao, R. Huang, W. Zhang, X. Jiang, et al. Wukong: A 100 million large-scale chinese cross-modal pre-training benchmark.Advances in Neural Information Processing Systems, 35:26418–26431, 2022

    work page 2022

  15. [15]

    T. Guan, F. Liu, X. Wu, R. Xian, Z. Li, X. Liu, X. Wang, L. Chen, F. Huang, Y . Yacoob, et al. Hallusionbench: an advanced diagnostic suite for entangled language hallucination and visual illusion in large vision-language models. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14375–14385, 2024

    work page 2024

  16. [16]

    D. Guo, F. Wu, F. Zhu, F. Leng, G. Shi, H. Chen, H. Fan, J. Wang, J. Jiang, J. Wang, et al. Seed1. 5-vl technical report.arXiv preprint arXiv:2505.07062, 2025

    work page internal anchor Pith review arXiv 2025

  17. [17]

    D. Guo, D. Yang, H. Zhang, J. Song, R. Zhang, R. Xu, Q. Zhu, S. Ma, P. Wang, X. Bi, et al. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning.arXiv preprint arXiv:2501.12948, 2025. 20

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  18. [18]

    H. He, W. Yao, K. Ma, W. Yu, Y . Dai, H. Zhang, Z. Lan, and D. Yu. Webvoyager: Building an end-to-end web agent with large multimodal models.arXiv preprint arXiv:2401.13919, 2024

    work page internal anchor Pith review arXiv 2024

  19. [19]

    Hong*, Y

    W. Hong*, Y . Cheng*, Z. Yang*, W. Wang, L. Wang, X. Gu, S. Huang, Y . Dong, and J. Tang. Motionbench: Benchmarking and improving fine-grained video motion understanding for vision language models, 2024

    work page 2024

  20. [20]

    W. Hong, W. Wang, M. Ding, W. Yu, Q. Lv, Y . Wang, Y . Cheng, S. Huang, J. Ji, Z. Xue, et al. Cogvlm2: Visual language models for image and video understanding.arXiv preprint arXiv:2408.16500, 2024

    work page internal anchor Pith review arXiv 2024

  21. [21]

    W. Hong, W. Wang, Q. Lv, J. Xu, W. Yu, J. Ji, Y . Wang, Z. Wang, Y . Dong, M. Ding, et al. Cogagent: A visual language model for gui agents. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14281–14290, 2024

    work page 2024

  22. [22]

    K. Hu, P. Wu, F. Pu, W. Xiao, Y . Zhang, X. Yue, B. Li, and Z. Liu. Video-mmmu: Evaluating knowledge acquisition from multi-discipline professional videos. 2025

    work page 2025

  23. [23]

    OpenAI o1 System Card

    A. Jaech, A. Kalai, A. Lerer, A. Richardson, A. El-Kishky, A. Low, A. Helyar, A. Madry, A. Beutel, A. Carney, et al. Openai o1 system card.arXiv preprint arXiv:2412.16720, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  24. [24]

    M. Jia, Z. Qi, S. Zhang, W. Zhang, X. Yu, J. He, H. Wang, and L. Yi. Omnispatial: To- wards comprehensive spatial reasoning benchmark for vision language models.arXiv preprint arXiv:2506.03135, 2025

    work page arXiv 2025

  25. [25]

    Kazemzadeh, V

    S. Kazemzadeh, V . Ordonez, M. Matten, and T. Berg. Referitgame: Referring to objects in photographs of natural scenes. InProceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), pages 787–798, 2014

    work page 2014

  26. [26]

    Kembhavi, M

    A. Kembhavi, M. Salvato, E. Kolve, M. Seo, H. Hajishirzi, and A. Farhadi. A diagram is worth a dozen images. InComputer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part IV 14, pages 235–251. Springer, 2016

    work page 2016

  27. [27]

    J. Li, D. Li, S. Savarese, and S. Hoi. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. InInternational conference on machine learning, pages 19730–19742. PMLR, 2023

    work page 2023

  28. [28]

    K. Li, Y . Wang, Y . He, Y . Li, Y . Wang, Y . Liu, Z. Wang, J. Xu, G. Chen, P. Luo, L. Wang, and Y . Qiao. MVBench: A comprehensive multi-modal video understanding benchmark, 2023

    work page 2023

  29. [29]

    Q. Li, Z. Chen, W. Wang, W. Wang, S. Ye, Z. Jin, G. Chen, Y . He, Z. Gao, E. Cui, et al. Omnicorpus: A unified multimodal corpus of 10 billion-level images interleaved with text. arXiv preprint arXiv:2406.08418, 2024

    work page arXiv 2024

  30. [30]

    Y . Liu, H. Duan, Y . Zhang, B. Li, S. Zhang, W. Zhao, Y . Yuan, J. Wang, C. He, Z. Liu, K. Chen, and D. Lin. Mmbench: Is your multi-modal model an all-around player?arXiv:2307.06281, 2023

    work page internal anchor Pith review arXiv 2023

  31. [31]

    Y . Liu, Z. Li, M. Huang, B. Yang, W. Yu, C. Li, X.-C. Yin, C.-L. Liu, L. Jin, and X. Bai. Ocrbench: on the hidden mystery of ocr in large multimodal models.Science China Information Sciences, 67(12), Dec. 2024

    work page 2024

  32. [32]

    P. Lu, H. Bansal, T. Xia, J. Liu, C. Li, H. Hajishirzi, H. Cheng, K.-W. Chang, M. Galley, and J. Gao. Mathvista: Evaluating mathematical reasoning of foundation models in visual contexts. arXiv preprint arXiv:2310.02255, 2023

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  33. [33]

    Y . Ma, L. Du, X. Shen, S. Chen, P. Li, Q. Ren, L. Ma, Y . Dai, P. Liu, and J. Yan. One rl to see them all: Visual triple unified reinforcement learning, 2025

    work page 2025

  34. [34]

    Y . Ma, Y . Zang, L. Chen, M. Chen, Y . Jiao, X. Li, X. Lu, Z. Liu, Y . Ma, X. Dong, P. Zhang, L. Pan, Y .-G. Jiang, J. Wang, Y . Cao, and A. Sun. Mmlongbench-doc: Benchmarking long- context document understanding with visualizations, 2024. 21

    work page 2024

  35. [35]

    Masry, M

    A. Masry, M. S. Islam, M. Ahmed, A. Bajaj, F. Kabir, A. Kartha, M. T. R. Laskar, M. Rahman, S. Rahman, M. Shahmohammadi, M. Thakkar, M. R. Parvez, E. Hoque, and S. Joty. Chartqapro: A more diverse and challenging benchmark for chart question answering, 2025

    work page 2025

  36. [36]

    OpenAI. Gpt-4o. 2024

    work page 2024

  37. [37]

    R. Qiao, Q. Tan, G. Dong, M. Wu, C. Sun, X. Song, Z. GongQue, S. Lei, Z. Wei, M. Zhang, et al. We-math: Does your large multimodal model achieve human-like mathematical reasoning? arXiv preprint arXiv:2407.01284, 2024

    work page internal anchor Pith review arXiv 2024

  38. [38]

    AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents

    C. Rawles, S. Clinckemaillie, Y . Chang, J. Waltz, G. Lau, M. Fair, A. Li, W. Bishop, W. Li, F. Campbell-Ajala, et al. Androidworld: A dynamic benchmarking environment for autonomous agents.arXiv:2405.14573, 2024

    work page internal anchor Pith review arXiv 2024

  39. [39]

    Jonathan Roberts, Mohammad Reza Taesiri, Ansh Sharma, Akash Gupta, Samuel Roberts, Ioana Croitoru, Simion-Vlad Bogolin, Jialu Tang, Flo- rian Langer, Vyas Raina, and 1 others

    J. Roberts, M. R. Taesiri, A. Sharma, A. Gupta, S. Roberts, I. Croitoru, S.-V . Bogolin, J. Tang, F. Langer, V . Raina, et al. Zerobench: An impossible visual benchmark for contemporary large multimodal models.arXiv preprint arXiv:2502.09696, 2025

    work page arXiv 2025

  40. [40]

    Schuhmann, R

    C. Schuhmann, R. Beaumont, R. Vencu, C. Gordon, R. Wightman, M. Cherti, T. Coombes, A. Katta, C. Mullis, M. Wortsman, et al. Laion-5b: An open large-scale dataset for training next generation image-text models.Advances in neural information processing systems, 35:25278– 25294, 2022

    work page 2022

  41. [41]

    R. Shao, S. S. Li, R. Xin, S. Geng, Y . Wang, S. Oh, S. S. Du, N. Lambert, S. Min, R. Krishna, et al. Spurious rewards: Rethinking training signals in rlvr.arXiv preprint arXiv:2506.10947, 2025

    work page internal anchor Pith review arXiv 2025

  42. [42]

    Z. Shao, P. Wang, Q. Zhu, R. Xu, J. Song, X. Bi, H. Zhang, M. Zhang, Y . Li, Y . Wu, et al. Deepseekmath: Pushing the limits of mathematical reasoning in open language models.arXiv preprint arXiv:2402.03300, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  43. [43]

    C. Si, Y . Zhang, Z. Yang, R. Liu, and D. Yang. Design2code: How far are we from automating front-end engineering?, 2024.URL https://arxiv. org/abs/2403, 3163, 2024

    work page 2024

  44. [44]

    J. Su, Y . Lu, S. Pan, A. Murtadha, B. Wen, and Y . Liu. Roformer: Enhanced transformer with rotary position embedding.arXiv preprint arXiv:2104.09864, 2021

    work page internal anchor Pith review Pith/arXiv arXiv 2021

  45. [45]

    L. Tang, G. Kim, X. Zhao, T. Lake, W. Ding, F. Yin, P. Singhal, M. Wadhwa, Z. L. Liu, Z. Sprague, et al. Chartmuseum: Testing visual reasoning capabilities of large vision-language models.arXiv preprint arXiv:2505.13444, 2025

    work page arXiv 2025

  46. [46]

    C. Team, Z. Yue, Z. Lin, Y . Song, W. Wang, S. Ren, S. Gu, S. Li, P. Li, L. Zhao, L. Li, K. Bao, H. Tian, H. Zhang, G. Wang, D. Zhu, Cici, C. He, B. Ye, B. Shen, Z. Zhang, Z. Jiang, Z. Zheng, Z. Song, Z. Luo, Y . Yu, Y . Wang, Y . Tian, Y . Tu, Y . Yan, Y . Huang, X. Wang, X. Xu, X. Song, X. Zhang, X. Yong, X. Zhang, X. Deng, W. Yang, W. Ma, W. Lv, W. Zhu...

    work page 2025

  47. [47]

    G. Team, R. Anil, S. Borgeaud, Y . Wu, J.-B. Alayrac, J. Yu, R. Soricut, J. Schalkwyk, A. M. Dai, A. Hauth, K. Millican, D. Silver, S. Petrov, M. Johnson, I. Antonoglou, J. Schrittwieser, A. Glaese, J. Chen, E. Pitler, T. Lillicrap, A. Lazaridou, O. Firat, J. Molloy, M. Isard, P. R. Barham, T. Hennigan, B. Lee, F. Viola, M. Reynolds, Y . Xu, R. Doherty, E...

    work page 2023

  48. [48]

    G. Team, A. Kamath, J. Ferret, S. Pathak, N. Vieillard, R. Merhej, S. Perrin, T. Matejovicova, A. Ramé, M. Rivière, et al. Gemma 3 technical report.arXiv preprint arXiv:2503.19786, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  49. [49]

    G. R. Team, S. Abeyruwan, J. Ainslie, J.-B. Alayrac, M. G. Arenas, T. Armstrong, A. Balakr- ishna, R. Baruch, M. Bauza, M. Blokzijl, et al. Gemini robotics: Bringing ai into the physical world.arXiv preprint arXiv:2503.20020, 2025

    work page internal anchor Pith review arXiv 2025

  50. [50]

    K. Team, A. Du, B. Yin, B. Xing, B. Qu, B. Wang, C. Chen, C. Zhang, C. Du, C. Wei, C. Wang, D. Zhang, D. Du, D. Wang, E. Yuan, E. Lu, F. Li, F. Sung, G. Wei, G. Lai, H. Zhu, H. Ding, H. Hu, H. Yang, H. Zhang, H. Wu, H. Yao, H. Lu, H. Wang, H. Gao, H. Zheng, J. Li, J. Su, J. Wang, J. Deng, J. Qiu, J. Xie, J. Wang, J. Liu, J. Yan, K. Ouyang, L. Chen, L. Sui...

    work page 2025

  51. [51]

    S. Tong, E. Brown, P. Wu, S. Woo, M. Middepogu, S. C. Akula, J. Yang, S. Yang, A. Iyer, X. Pan, A. Wang, R. Fergus, Y . LeCun, and S. Xie. Cambrian-1: A fully open, vision-centric exploration of multimodal llms, 2024

    work page 2024

  52. [52]

    B. Wang, B. Wang, C. Wan, G. Huang, H. Hu, H. Jia, H. Nie, M. Li, N. Chen, S. Chen, et al. Step-3 is large yet affordable: Model-system co-design for cost-effective decoding.arXiv preprint arXiv:2507.19427, 2025

    work page arXiv 2025

  53. [53]

    F. Wang, X. Fu, J. Y . Huang, Z. Li, Q. Liu, X. Liu, M. D. Ma, N. Xu, W. Zhou, K. Zhang, et al. Muirbench: A comprehensive benchmark for robust multi-image understanding.arXiv preprint arXiv:2406.09411, 2024

    work page internal anchor Pith review arXiv 2024

  54. [54]

    H. Wang, X. Li, Z. Huang, A. Wang, J. Wang, T. Zhang, J. Zheng, S. Bai, Z. Kang, J. Feng, et al. Traceable evidence enhanced visual grounded reasoning: Evaluation and methodology. arXiv preprint arXiv:2507.07999, 2025

    work page arXiv 2025

  55. [55]

    K. Wang, J. Pan, W. Shi, Z. Lu, M. Zhan, and H. Li. Measuring multimodal mathematical reasoning with math-vision dataset.arXiv:2402.14804, 2024

    work page internal anchor Pith review arXiv 2024

  56. [56]

    M. Wang, S. Sunkara, G. Baechler, J. Lin, Y . Zhu, F. Zubach, L. Shu, and J. Chen. Webquest: A benchmark for multimodal qa on web page sequences, 2024

    work page 2024

  57. [57]

    P. Wang, S. Bai, S. Tan, S. Wang, Z. Fan, J. Bai, K. Chen, X. Liu, J. Wang, W. Ge, et al. Qwen2-vl: Enhancing vision-language model’s perception of the world at any resolution.arXiv preprint arXiv:2409.12191, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  58. [58]

    W. Wang, Z. He, W. Hong, Y . Cheng, X. Zhang, J. Qi, S. Huang, B. Xu, Y . Dong, M. Ding, et al. Lvbench: An extreme long video understanding benchmark.arXiv preprint arXiv:2406.08035, 2024

    work page internal anchor Pith review arXiv 2024

  59. [59]

    W. Wang, Q. Lv, W. Yu, W. Hong, J. Qi, Y . Wang, J. Ji, Z. Yang, L. Zhao, X. Song, et al. Cogvlm: Visual expert for pretrained language models.arXiv preprint arXiv:2311.03079, 2023

    work page internal anchor Pith review arXiv 2023

  60. [60]

    J. Wei, X. Wang, D. Schuurmans, M. Bosma, B. Ichter, F. Xia, E. H. Chi, Q. V . Le, and D. Zhou. Chain-of-thought prompting elicits reasoning in large language models. In S. Koyejo, S. Mohamed, A. Agarwal, D. Belgrave, K. Cho, and A. Oh, editors,Proc. of Neural Information Processing Systems, 2022

    work page 2022

  61. [61]

    Y . Xiao, E. Sun, T. Liu, and W. Wang. Logicvista: Multimodal llm logical reasoning benchmark in visual contexts, 2024

    work page 2024

  62. [62]

    T. Xie, D. Zhang, J. Chen, X. Li, S. Zhao, R. Cao, J. H. Toh, Z. Cheng, D. Shin, F. Lei, et al. Osworld: Benchmarking multimodal agents for open-ended tasks in real computer environments. Advances in Neural Information Processing Systems, 37:52040–52094, 2025

    work page 2025

  63. [63]

    H. Xu, S. Xie, X. E. Tan, P.-Y . Huang, R. Howes, V . Sharma, S.-W. Li, G. Ghosh, L. Zettlemoyer, and C. Feichtenhofer. Demystifying clip data.arXiv preprint arXiv:2309.16671, 2023

    work page internal anchor Pith review arXiv 2023

  64. [64]

    Y . Xu, H. Dong, L. Wang, D. Sahoo, J. Li, and C. Xiong. Scalable chain of thoughts via elastic reasoning.arXiv preprint arXiv:2505.05315, 2025

    work page arXiv 2025

  65. [65]

    C.-H. Yeh, C. Wang, S. Tong, T.-Y . Cheng, R. Wang, T. Chu, Y . Zhai, Y . Chen, S. Gao, and Y . Ma. Seeing from another perspective: Evaluating multi-view understanding in mllms.arXiv preprint arXiv:2504.15280, 2025

    work page arXiv 2025

  66. [66]

    Q. Yu, Z. Zhang, R. Zhu, Y . Yuan, X. Zuo, Y . Yue, W. Dai, T. Fan, G. Liu, L. Liu, et al. Dapo: An open-source llm reinforcement learning system at scale.arXiv preprint arXiv:2503.14476, 2025. 25

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  67. [67]

    X. Yue, Y . Ni, K. Zhang, T. Zheng, R. Liu, G. Zhang, S. Stevens, D. Jiang, W. Ren, Y . Sun, C. Wei, B. Yu, R. Yuan, R. Sun, M. Yin, B. Zheng, Z. Yang, Y . Liu, W. Huang, H. Sun, Y . Su, and W. Chen. Mmmu: A massive multi-discipline multimodal understanding and reasoning benchmark for expert agi. InProc. of Computer Vision and Pattern Recognition, 2024

    work page 2024

  68. [68]

    X. Yue, T. Zheng, Y . Ni, Y . Wang, K. Zhang, S. Tong, Y . Sun, B. Yu, G. Zhang, H. Sun, Y . Su, W. Chen, and G. Neubig. Mmmu-pro: A more robust multi-discipline multimodal understanding benchmark.arXiv preprint arXiv:2409.02813, 2024

    work page internal anchor Pith review arXiv 2024

  69. [69]

    Zhang, P

    H. Zhang, P. Zhang, X. Hu, Y .-C. Chen, L. Li, X. Dai, L. Wang, L. Yuan, J.-N. Hwang, and J. Gao. Glipv2: Unifying localization and vision-language understanding.Proc. of Neural Information Processing Systems, 35:36067–36080, 2022

    work page 2022

  70. [70]

    Zhang, D

    R. Zhang, D. Jiang, Y . Zhang, H. Lin, Z. Guo, P. Qiu, A. Zhou, P. Lu, K.-W. Chang, Y . Qiao, et al. Mathverse: Does your multi-modal llm truly see the diagrams in visual math problems? InEuropean Conference on Computer Vision, pages 169–186. Springer, 2024

    work page 2024

  71. [71]

    Y . Zhao, L. Xie, H. Zhang, G. Gan, Y . Long, Z. Hu, T. Hu, W. Chen, C. Li, J. Song, Z. Xu, C. Wang, W. Pan, Z. Shangguan, X. Tang, Z. Liang, Y . Liu, C. Zhao, and A. Cohan. Mmvu: Measuring expert-level multi-discipline video understanding, 2025

    work page 2025

  72. [72]

    J. Zhu, W. Wang, Z. Chen, Z. Liu, S. Ye, L. Gu, H. Tian, Y . Duan, W. Su, J. Shao, et al. Internvl3: Exploring advanced training and test-time recipes for open-source multimodal models.arXiv preprint arXiv:2504.10479, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  73. [73]

    W. Zhu, J. Hessel, A. Awadalla, S. Y . Gadre, J. Dodge, A. Fang, Y . Yu, L. Schmidt, W. Y . Wang, and Y . Choi. Multimodal c4: An open, billion-scale corpus of images interleaved with text. Advances in Neural Information Processing Systems, 36:8958–8974, 2023

    work page 2023

  74. [74]

    Dynamath: A dynamic visual benchmark for evaluating mathematical reasoning robustness of vision language models.arXiv preprint arXiv:2411.00836, 2024

    C. Zou, X. Guo, R. Yang, J. Zhang, B. Hu, and H. Zhang. Dynamath: A dynamic visual benchmark for evaluating mathematical reasoning robustness of vision language models.arXiv preprint arXiv:2411.00836, 2024. 26 A Qualitative Examples In this section, we demonstrate the capabilities and advantages of our model through various qualita- tive examples. A.1 UI ...

    work page arXiv 2024

  75. [75]

    毛细管”。 当左右两个装有不同颜色液体的杯子与中间的空杯之间用纸巾连接时，纸巾会利用自身吸水性和纤维间的毛细作 用，将左侧红色液体和右侧蓝色液体通过纤维间隙输送至中间的空杯中。随着这种输送过程的进行，中间的空杯逐 渐被液体填满，从而出现了“中间水杯有水

    识别实验装置：三个玻璃杯，左右装有色液体，中间空，中间有纸巾连接。2. 确定物理原理：毛细现象（或毛细 作用），纸巾的多孔性使其能吸附液体并通过纤维间的通道输送。3. 解释过程：纸巾吸收左侧红色液体和右侧蓝色 液体，通过毛细作用将这些液体输送到中间的空杯中，因此中间水杯里有水（实际上是两种液体通过纸巾的毛细作 用流入中间杯）。 所以答案要说明是毛细现象（或毛细作用），纸巾的吸水和导水能力导致中间杯有水。</think><answer>这个现 象是由于**毛细现象（或毛细作用）**导致的。具体来说：纸巾是多孔材料，其纤维之间的缝隙类似于“毛细管”。 当左右两个装有不同颜色液体的杯子与中间的空杯之间用纸巾连接时，纸巾会利用自身吸水性和纤维间的毛细作 用，将左侧红色液体和右侧蓝色液体通过纤维间隙输送至中间的空杯...

    work page

  76. [76]

    Meeting" event - October 9th has a

    The calendar displays dates from October 1-31, with days of the week labeled as S, M, T, W, T, F, S at the top. 3. Today appears to be October 15th, as indicated by the blue highlighted circle around the number 15 in the fourth row. 4. The task requires creating an event in two weeks from today (October 15th) which would be October 29th. 5. Looking at the...

    work page

  77. [77]

    GLM-4.1V-9B-Thinking sets a new state-of-the-art across 23 out of 28 benchmarks among models under 10B parameters, including MiMo-VL [ 46], Kimi-VL [50], InternVL3 [72], and Qwen-VL series [4]. 41 Task Benchmark GLM-4.1V-9B-ThinkingQwen2.5-VL7B InternVL39B Kimi-VLA3B-ThinkingMiMo-VL7B-RLQwen2.5-VL72B GPT-4o2024-11-20 General VQA MMBench-V1.1-EN85.8 82.7 8...

    work page