1
面壁智能发布新一代端侧多模态大模型 MiniCPM-V 4.7，仅 3GB 内存可实现 13+ token/s 解码速度
IT之家 10 月 13 日消息，面壁智能今天宣布推出新一代端侧多模态大模型 MiniCPM-V 4.7，提供 35B-A3B 和 1B 两种规格，其中 35B-A3B 是 MiniCPM-V 系列首个 MoE 模型。 据官方介绍，这款模型基于最新 LES（前瞻专家调度）推理加速技术，其中 35B-…
2
MedReaMM: Evaluating Large Multimodal Models on Expert-Level Clinical Diagnostic Synthesis
破解AI看病难题：全新基准MedReaMM如何考核多模态大模型的专业诊断整合能力
arXiv:2608.22323v1 Announce Type: new Abstract: The application of Large Language Models (LLMs) to diagnostic decision-making has garnered growing int…
3
Beyond RGB: Benchmarking and Enhancing MLLMs for Hyperspectral Image Understanding via Training-Free Reasoning Framework
多模态大模型不止看懂RGB，高光谱成像理解迎来全新基准与免训练推理框架。
arXiv:2604.08884v2 Announce Type: replace-cross Abstract: Multimodal Large Language Models (MLLMs) have achieved strong performance on RGB image under…
4
Anchoring Bias: A Persistent Fairness Backdoor Attack against MLLMs under Continual Learning
揭秘多模态大模型在持续学习中的公平性后门攻击，锚定偏差如何悄然埋下隐患。
arXiv:2608.21577v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) are increasingly deployed in high-stakes domains where fair…
5
Context-Aware Cluster Decoding: Semantic Anchor-Driven Coherence in dMLLMs
多模态大模型解码新思路：用语义锚点驱动聚类生成，让长文本输出更连贯，EMNLP 2026 已收录，值得跟进。
arXiv:2608.22367v1 Announce Type: new Abstract: Diffusion multimodal large language models (dMLLMs) frequently produce long-form outputs marred by sem…
6
Recognition-Conditioned Reasoning: A Training-Free Multimodal-LLM Pipeline for Fine-Grained Micro-Action Understanding
无需训练的多模态大模型流水线，精准解析微动作语义，ACM Multimedia 2026亮点工作
arXiv:2608.21022v1 Announce Type: new Abstract: Micro-actions are subtle, short, low-amplitude body movements, such as a fidgeting hand or a slight he…
7
原生支持 4K 图像输出，商汤开源 8B 参数量多模态大模型 SenseNova U1.5 Lite
商汤开源8B多模态大模型，原生4K图像输出，轻量却媲美超大规模商业模型
IT之家 8 月 21 日消息，商汤科技今日宣布 开源轻量级、原生统一 多模态 大模型 SenseNova U1.5 Lite。 据介绍，SenseNova U1.5 Lite 原生支持 3-4k 上下文长度 ，能够同时处理主体、数量、空间关系、文字、布局、风格等多重约束，提升复杂视觉任务的执行稳定…
8
MR-IQA-2: Faithful Image Quality Reflection via Fine-Grained Credit Assignment
提出MR-IQA-2，用细粒度信用分配让多模态大模型在图像质量评估中的推理更忠实。
arXiv:2608.18579v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) have shown strong potential for image quality assessment (I…
9
PATE-Forensics: Perception-as-Tool for Explainable Deepfake Forensics with General-Purpose MLLMs
用通用多模态大模型把深度伪造检测变成可解释的感知工具，让AI“说”出造假痕迹。
arXiv:2608.18573v1 Announce Type: new Abstract: Existing explainable deepfake forensic methods typically rely on task-adapted MLLM to jointly address …
10
Uncertainty-Aware Decision Making in Multimodal Large Language Models
多模态大模型如何量化不确定性？这项研究为高风险场景下的可靠决策提供新思路。
arXiv:2608.17084v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) increasingly answer questions whose correctness depends on vi…
11
MLLM-Guided Semantic Correction for Text-to-Video Generation
多模态大模型当导演，精准修正文本到视频的语义偏差，生成更贴合提示的视频。
arXiv:2608.16513v1 Announce Type: cross Abstract: Recent advances in diffusion models and Transformer architectures have led to significant progress i…
12
Remote-Sensing City Layout Extraction with MLLM
用多模态大模型把遥感影像直接变成可执行的程序化城市布局，打通识别到重建的闭环。
arXiv:2608.16484v1 Announce Type: new Abstract: Remote-sensing systems usually describe urban content with detection boxes, semantic masks, or vector …
13
Diagram-MMU: A Multi-Modal Benchmark for Scientific Diagrams
科学图表如何变成LaTeX代码？新基准直击多模态大模型短板。
arXiv:2608.12262v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) have been growing the capability for scientific writing and…
14
LEGO-Puzzles: How Good Are MLLMs at Multi-Step Spatial Reasoning?
乐高拼图当考题，多模态大模型的多步空间推理能力一测便知，机器人、自动驾驶都靠它。
arXiv:2503.19990v4 Announce Type: replace Abstract: Many real-world applications of spatial intelligence, such as robotic control, autonomous driving,…
15
NeuPAT: Neuron-aware Plasticity Allocation Tuning for Language-Preserving MLLMs
多模态大模型语言能力易退化？神经元感知的可塑性分配调优带来新解法。
arXiv:2608.08107v1 Announce Type: cross Abstract: Multimodal expansion of large language models (LLMs) enables new perceptual capabilities but often c…
16
An AI4AI Framework for Visual Token Pruning
把Token剪枝交给AI自动设计，告别手工调参，多模态大模型推理成本有望再降一截。
arXiv:2608.07193v1 Announce Type: new Abstract: Visual-token pruning can substantially reduce the inference cost of multimodal large language models (…
17
RoRA: Role-Oriented Regional Allocation for Visual Token Pruning in MLLMs
多模态大模型提速新思路：按角色分区裁剪视觉token，兼顾效率与精度
arXiv:2608.07088v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) encode images as long visual token sequences, making prefil…
18
From Sports to Safety: Benchmarking Proactive Risk Inference in MLLMs
多模态大模型能否预判危险？这份基准测试从运动场景到安全风险，评测AI的主动风险推断能力。
arXiv:2608.05560v1 Announce Type: cross Abstract: Timely anticipation of physical hazards is essential for real-world safety, yet existing MLLM evalua…
19
Beyond Frame Selection: Rethinking Long-Video Understanding with MLLMs
跳出帧采样旧范式，重新思考多模态大模型如何真正理解长视频。
arXiv:2608.05592v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) have achieved strong progress in video understanding, yet it …
20
Seeing Is Not Deciding: Can Multimodal LLMs Act as Effective CEOs?
多模态大模型能否胜任CEO决策？这篇论文用实验揭示视觉信息与决策能力之间的鸿沟，视角新颖。
arXiv:2608.05864v1 Announce Type: new Abstract: Large language models are increasingly applied as autonomous decision-making agents. However, in execu…