Guanzhou Chen

Zihao Wang, Shaofei Cai, Guanzhou Chen et al. · pku

We investigate the challenge of task planning for multi-task embodied agents in open-world environments. Two main difficulties are identified: 1) executing plans in an open-world environment (e.g., Minecraft) necessitates accurate and multi-step reasoning due to the long-term nature of tasks, and 2) as vanilla planners do not consider how easy the current agent can achieve a given sub-task when ordering parallel sub-goals within a complicated plan, the resulting plan could be inefficient or even infeasible. To this end, we propose "$\underline{D}$escribe, $\underline{E}$xplain, $\underline{P}$lan and $\underline{S}$elect" ($\textbf{DEPS}$), an interactive planning approach based on Large Language Models (LLMs). DEPS facilitates better error correction on initial LLM-generated $\textit{plan}$ by integrating $\textit{description}$ of the plan execution process and providing self-$\textit{explanation}$ of feedback when encountering failures during the extended planning phases. Furthermore, it includes a goal $\textit{selector}$, which is a trainable module that ranks parallel candidate sub-goals based on the estimated steps of completion, consequently refining the initial plan. Our experiments mark the milestone of the first zero-shot multi-task agent that can robustly accomplish 70+ Minecraft tasks and nearly double the overall performances. Further testing reveals our method's general effectiveness in popularly adopted non-open-ended domains as well (i.e., ALFWorld and tabletop manipulation). The ablation and exploratory studies detail how our design beats the counterparts and provide a promising update on the $\texttt{ObtainDiamond}$ grand challenge with our approach. The code is released at https://github.com/CraftJarvis/MC-Planner.

Yicheng Zou, Dongsheng Zhu, Lin Zhu et al.

We introduce Intern-S1-Pro, the first one-trillion-parameter scientific multimodal foundation model. Scaling to this unprecedented size, the model delivers a comprehensive enhancement across both general and scientific domains. Beyond stronger reasoning and image-text understanding capabilities, its intelligence is augmented with advanced agent capabilities. Simultaneously, its scientific expertise has been vastly expanded to master over 100 specialized tasks across critical science fields, including chemistry, materials, life sciences, and earth sciences. Achieving this massive scale is made possible by the robust infrastructure support of XTuner and LMDeploy, which facilitates highly efficient Reinforcement Learning (RL) training at the 1-trillion parameter level while ensuring strict precision consistency between training and inference. By seamlessly integrating these advancements, Intern-S1-Pro further fortifies the fusion of general and specialized intelligence, working as a Specializable Generalist, demonstrating its position in the top tier of open-source models for general capabilities, while outperforming proprietary models in the depth of specialized scientific tasks.

Yibo Yang, Fei Lei, Yixuan Sun et al.

As AI-driven document understanding and processing tools become increasingly prevalent in real-world applications, the need for rigorous evaluation standards has grown increasingly urgent. Existing benchmarks and evaluations often focus on isolated capabilities or simplified scenarios, failing to capture the end-to-end task effectiveness required in practical settings. To address this gap, we introduce AIDABench, a comprehensive benchmark for evaluating AI systems on complex data analytics tasks in an end-to-end manner. AIDABench encompasses 600+ diverse document analysis tasks across three core capability dimensions: question answering, data visualization, and file generation. These tasks are grounded in realistic scenarios involving heterogeneous data types, including spreadsheets, databases, financial reports, and operational records, and reflect analytical demands across diverse industries and job functions. Notably, the tasks in AIDABench are sufficiently challenging that even human experts require 1-2 hours per question when assisted by AI tools, underscoring the benchmark's difficulty and real-world complexity. We evaluate 11 state-of-the-art models on AIDABench, spanning both proprietary (e.g., Claude Sonnet 4.5, Gemini 3 Pro Preview) and open-source (e.g., Qwen3-Max-2026-01-23-Thinking) families. Our results reveal that complex, real-world data analytics tasks remain a significant challenge for current AI systems, with the best-performing model achieving only 59.43% pass-at-1. We provide a detailed analysis of failure modes across each capability dimension and identify key challenges for future research. AIDABench offers a principled reference for enterprise procurement, tool selection, and model optimization, and is publicly available at https://github.com/MichaelYang-lyx/AIDABench.

Guanzhou Chen, Erfei Cui, Changyao Tian et al.

Instruction-based image editing has emerged as a key capability for unified multimodal models (UMMs), yet constructing large-scale, diverse, and high-quality editing datasets without costly proprietary APIs remains challenging. Previous image editing datasets either rely on closed-source models for annotation, which prevents cost-effective scaling, or employ fixed synthetic editing pipelines, which suffer from limited quality and generalizability. To address these challenges, we propose ScaleEditor, a fully open-source hierarchical multi-agent framework for end-to-end construction of large-scale, high-quality image editing datasets. Our pipeline consists of three key components: source image expansion with world-knowledge infusion, adaptive multi-agent editing instruction-image synthesis, and a task-aware data quality verification mechanism. Using ScaleEditor, we curate ScaleEdit-12M, the largest open-source image editing dataset to date, spanning 23 task families across diverse real and synthetic domains. Fine-tuning UniWorld-V1 and Bagel on ScaleEdit yields consistent gains, improving performance by up to 10.4% on ImgEdit and 35.1% on GEdit for general editing benchmarks and by up to 150.0% on RISE and 26.5% on KRIS-Bench for knowledge-infused benchmarks. These results demonstrate that open-source, agentic pipelines can approach commercial-grade data quality while retaining cost-effectiveness and scalability. Both the framework and dataset will be open-sourced.

Yangguang Chen, Tong Wang, Guanzhou Chen et al.

The detection of façade elements on buildings, such as doors, windows, balconies, air conditioning units, billboards, and glass curtain walls, is a critical step in automating the creation of Building Information Modeling (BIM). Yet, this field faces significant challenges, including the uneven distribution of façade elements, the presence of small objects, and substantial background noise, which hamper detection accuracy. To address these issues, we develop the BFA-YOLO model and the BFA-3D dataset in this study. The BFA-YOLO model is an advanced architecture designed specifically for analyzing multi-view images of façade attachments. It integrates three novel components: the Feature Balanced Spindle Module (FBSM) that tackles the issue of uneven object distribution; the Target Dynamic Alignment Task Detection Head (TDATH) that enhances the detection of small objects; and the Position Memory Enhanced Self-Attention Mechanism (PMESA), aimed at reducing the impact of background noise. These elements collectively enable BFA-YOLO to effectively address each challenge, thereby improving model robustness and detection precision. The BFA-3D dataset, offers multi-view images with precise annotations across a wide range of façade attachment categories. This dataset is developed to address the limitations present in existing façade detection datasets, which often feature a single perspective and insufficient category coverage. Through comparative analysis, BFA-YOLO demonstrated improvements of 1.8\% and 2.9\% in mAP$_{50}$ on the BFA-3D dataset and the public Façade-WHU dataset, respectively, when compared to the baseline YOLOv8 model. These results highlight the superior performance of BFA-YOLO in façade element detection and the advancement of intelligent BIM technologies.

Weiyun Wang, Zhangwei Gao, Lixin Gu et al. · cmu, pku

We introduce InternVL 3.5, a new family of open-source multimodal models that significantly advances versatility, reasoning capability, and inference efficiency along the InternVL series. A key innovation is the Cascade Reinforcement Learning (Cascade RL) framework, which enhances reasoning through a two-stage process: offline RL for stable convergence and online RL for refined alignment. This coarse-to-fine training strategy leads to substantial improvements on downstream reasoning tasks, e.g., MMMU and MathVista. To optimize efficiency, we propose a Visual Resolution Router (ViR) that dynamically adjusts the resolution of visual tokens without compromising performance. Coupled with ViR, our Decoupled Vision-Language Deployment (DvD) strategy separates the vision encoder and language model across different GPUs, effectively balancing computational load. These contributions collectively enable InternVL3.5 to achieve up to a +16.0\% gain in overall reasoning performance and a 4.05$\times$ inference speedup compared to its predecessor, i.e., InternVL3. In addition, InternVL3.5 supports novel capabilities such as GUI interaction and embodied agency. Notably, our largest model, i.e., InternVL3.5-241B-A28B, attains state-of-the-art results among open-source MLLMs across general multimodal, reasoning, text, and agentic tasks -- narrowing the performance gap with leading commercial models like GPT-5. All models and code are publicly released.

Changyao Tian, Danni Yang, Guanzhou Chen et al.

Unified multimodal models (UMMs) that integrate understanding, reasoning, generation, and editing face inherent trade-offs between maintaining strong semantic comprehension and acquiring powerful generation capabilities. In this report, we present InternVL-U, a lightweight 4B-parameter UMM that democratizes these capabilities within a unified framework. Guided by the principles of unified contextual modeling and modality-specific modular design with decoupled visual representations, InternVL-U integrates a state-of-the-art Multimodal Large Language Model (MLLM) with a specialized MMDiT-based visual generation head. To further bridge the gap between aesthetic generation and high-level intelligence, we construct a comprehensive data synthesis pipeline targeting high-semantic-density tasks, such as text rendering and scientific reasoning, under a reasoning-centric paradigm that leverages Chain-of-Thought (CoT) to better align abstract user intent with fine-grained visual generation details. Extensive experiments demonstrate that InternVL-U achieves a superior performance - efficiency balance. Despite using only 4B parameters, it consistently outperforms unified baseline models with over 3x larger scales such as BAGEL (14B) on various generation and editing tasks, while retaining strong multimodal understanding and reasoning capabilities.

Xiaoliang Tan, Guanzhou Chen, Tong Wang et al.

The field of Remote Sensing (RS) widely employs Change Detection (CD) on very-high-resolution (VHR) images. A majority of extant deep-learning-based methods hinge on annotated samples to complete the CD process. Recently, the emergence of Vision Foundation Model (VFM) enables zero-shot predictions in particular vision tasks. In this work, we propose an unsupervised CD method named Segment Change Model (SCM), built upon the Segment Anything Model (SAM) and Contrastive Language-Image Pre-training (CLIP). Our method recalibrates features extracted at different scales and integrates them in a top-down manner to enhance discriminative change edges. We further design an innovative Piecewise Semantic Attention (PSA) scheme, which can offer semantic representation without training, thereby minimize pseudo change phenomenon. Through conducting experiments on two public datasets, the proposed SCM increases the mIoU from 46.09% to 53.67% on the LEVIR-CD dataset, and from 47.56% to 52.14% on the WHU-CD dataset. Our codes are available at https://github.com/StephenApX/UCD-SCM.

Qingyuan Yang, Guanzhou Chen, Xiaoliang Tan et al.

Stereo matching and semantic segmentation are significant tasks in binocular satellite 3D reconstruction. However, previous studies primarily view these as independent parallel tasks, lacking an integrated multitask learning framework. This work introduces a solution, the Single-branch Semantic Stereo Network (S3Net), which innovatively combines semantic segmentation and stereo matching using Self-Fuse and Mutual-Fuse modules. Unlike preceding methods that utilize semantic or disparity information independently, our method dentifies and leverages the intrinsic link between these two tasks, leading to a more accurate understanding of semantic information and disparity estimation. Comparative testing on the US3D dataset proves the effectiveness of our S3Net. Our model improves the mIoU in semantic segmentation from 61.38 to 67.39, and reduces the D1-Error and average endpoint error (EPE) in disparity estimation from 10.051 to 9.579 and 1.439 to 1.403 respectively, surpassing existing competitive methods. Our codes are available at:https://github.com/CVEO/S3Net.

Zhenxin Lei, Zhangwei Gao, Changyao Tian et al.

Generalist visual captioning goes beyond a simple appearance description task, but requires integrating a series of visual cues into a caption and handling various visual domains. In this task, current open-source models present a large performance gap with commercial ones, which limits various applications such as data synthesis. To bridge the gap, this paper proposes CapFlow, a novel multi-agent collaboration workflow. CapFlow demonstrates for the first time that, by capitalizing on open-source models, it is possible to achieve caption quality on par with GPT-4.1 in various domains with an 89.5% reduction in costs. By leveraging CapFlow as the data synthesizer, we produce high-quality visual captions from image and video domains at scale, and obtain a generalist visual captioner via fine-tuning, namely MetaCaptioner. Through extensive experiments, we show that MetaCaptioner not only achieves comparable captioning capabilities with commercial models but also reaches top-tier multimodal performance in the open-source community. We hope CapFlow and MetaCaptioner can benefit future multimodal research by providing a strong and cost-effective visual captioning solution.

Tong Wang, Guanzhou Chen, Xiaodong Zhang et al.

Remote sensing image interpretation plays a critical role in environmental monitoring, urban planning, and disaster assessment. However, acquiring high-quality labeled data is often costly and time-consuming. To address this challenge, we proposes a multi-modal self-supervised learning framework that leverages high-resolution RGB images, multi-spectral data, and digital surface models (DSM) for pre-training. By designing an information-aware adaptive masking strategy, cross-modal masking mechanism, and multi-task self-supervised objectives, the framework effectively captures both the correlations across different modalities and the unique feature structures within each modality. We evaluated the proposed method on multiple downstream tasks, covering typical remote sensing applications such as scene classification, semantic segmentation, change detection, object detection, and depth estimation. Experiments are conducted on 15 remote sensing datasets, encompassing 26 tasks. The results demonstrate that the proposed method outperforms existing pretraining approaches in most tasks. Specifically, on the Potsdam and Vaihingen semantic segmentation tasks, our method achieved mIoU scores of 78.30\% and 76.50\%, with only 50\% train-set. For the US3D depth estimation task, the RMSE error is reduced to 0.182, and for the binary change detection task in SECOND dataset, our method achieved mIoU scores of 47.51\%, surpassing the second CS-MAE by 3 percentage points. Our pretrain code, checkpoints, and HR-Pairs dataset can be found in https://github.com/CVEO/MSSDF.

Qingyun Li, Zhe Chen, Weiyun Wang et al.

Image-text interleaved data, consisting of multiple images and texts arranged in a natural document format, aligns with the presentation paradigm of internet data and closely resembles human reading habits. Recent studies have shown that such data aids multimodal in-context learning and maintains the capabilities of large language models during multimodal fine-tuning. However, the limited scale and diversity of current image-text interleaved data restrict the development of multimodal large language models. In this paper, we introduce OmniCorpus, a 10 billion-scale image-text interleaved dataset. Using an efficient data engine, we filter and extract large-scale high-quality documents, which contain 8.6 billion images and 1,696 billion text tokens. Compared to counterparts (e.g., MMC4, OBELICS), our dataset 1) has 15 times larger scales while maintaining good data quality; 2) features more diverse sources, including both English and non-English websites as well as video-centric websites; 3) is more flexible, easily degradable from an image-text interleaved format to pure text corpus and image-text pairs. Through comprehensive analysis and experiments, we validate the quality, usability, and effectiveness of the proposed dataset. We hope this could provide a solid data foundation for future multimodal model research. Code and data are released at https://github.com/OpenGVLab/OmniCorpus.

Tong Wang, Guanzhou Chen, Xiaodong Zhang et al.

Despite the rapid evolution of semantic segmentation for land cover classification in high-resolution remote sensing imagery, integrating multiple data modalities such as Digital Surface Model (DSM), RGB, and Near-infrared (NIR) remains a challenge. Current methods often process only two types of data, missing out on the rich information that additional modalities can provide. Addressing this gap, we propose a novel \textbf{L}ightweight \textbf{M}ultimodal data \textbf{F}usion \textbf{Net}work (LMFNet) to accomplish the tasks of fusion and semantic segmentation of multimodal remote sensing images. LMFNet uniquely accommodates various data types simultaneously, including RGB, NirRG, and DSM, through a weight-sharing, multi-branch vision transformer that minimizes parameter count while ensuring robust feature extraction. Our proposed multimodal fusion module integrates a \textit{Multimodal Feature Fusion Reconstruction Layer} and \textit{Multimodal Feature Self-Attention Fusion Layer}, which can reconstruct and fuse multimodal features. Extensive testing on public datasets such as US3D, ISPRS Potsdam, and ISPRS Vaihingen demonstrates the effectiveness of LMFNet. Specifically, it achieves a mean Intersection over Union ($mIoU$) of 85.09\% on the US3D dataset, marking a significant improvement over existing methods. Compared to unimodal approaches, LMFNet shows a 10\% enhancement in $mIoU$ with only a 0.5M increase in parameter count. Furthermore, against bimodal methods, our approach with trilateral inputs enhances $mIoU$ by 0.46 percentage points.

Gen Luo, Ganlin Yang, Ziyang Gong et al.

The remarkable progress of Multimodal Large Language Models (MLLMs) has attracted increasing attention to extend them to physical entities like legged robot. This typically requires MLLMs to not only grasp multimodal understanding abilities, but also integrate visual-spatial reasoning and physical interaction capabilities. Nevertheless,existing methods struggle to unify these capabilities due to their fundamental differences.In this paper, we present the Visual Embodied Brain (VeBrain), a unified framework for perception, reasoning, and control in real world. VeBrain reformulates robotic control into common text-based MLLM tasks in the 2D visual space, thus unifying the objectives and mapping spaces of different tasks. Then, a novel robotic adapter is proposed to convert textual control signals from MLLMs to motion policies of real robots. From the data perspective, we further introduce VeBrain-600k, a high-quality instruction dataset encompassing various capabilities of VeBrain. In VeBrain-600k, we take hundreds of hours to collect, curate and annotate the data, and adopt multimodal chain-of-thought(CoT) to mix the different capabilities into a single conversation. Extensive experiments on 13 multimodal benchmarks and 5 spatial intelligence benchmarks demonstrate the superior performance of VeBrain to existing MLLMs like Qwen2.5-VL. When deployed to legged robots and robotic arms, VeBrain shows strong adaptability, flexibility, and compositional capabilities compared to existing methods. For example, compared to Qwen2.5-VL, VeBrain not only achieves substantial gains on MMVet by +5.6%, but also excels in legged robot tasks with +50% average gains.

Ganlin Yang, Tianyi Zhang, Haoran Hao et al.

While significant research has focused on developing embodied reasoning capabilities using Vision-Language Models (VLMs) or integrating advanced VLMs into Vision-Language-Action (VLA) models for end-to-end robot control, few studies directly address the critical gap between upstream VLM-based reasoning and downstream VLA policy learning. In this work, we take an initial step toward bridging embodied reasoning with VLA policy learning by introducing Vlaser - a Vision-Language-Action Model with synergistic embodied reasoning capability, which is a foundational vision-language model designed to integrate high-level reasoning with low-level control for embodied agents. Built upon the high-quality Vlaser-6M dataset, Vlaser achieves state-of-the-art performance across a range of embodied reasoning benchmarks - including spatial reasoning, embodied grounding, embodied QA, and task planning. Furthermore, we systematically examine how different VLM initializations affect supervised VLA fine-tuning, offering novel insights into mitigating the domain shift between internet-scale pre-training data and embodied-specific policy learning data. Based on these insights, our approach achieves state-of-the-art results on the WidowX benchmark and competitive performance on the Google Robot benchmark.

Xiangxu Wang, Tianhong Zhao, Wei Tu et al.

Origin-Destination (OD) flow matrices are essential for urban mobility analysis, underpinning applications in traffic forecasting, infrastructure planning, and policy design. However, existing methods suffer from two critical limitations: (1) reliance on auxiliary features (e.g., Points of Interest, socioeconomic statistics) that are costly to collect and have limited spatial coverage; and (2) sensitivity to spatial topology, where minor index reordering of urban regions (e.g., census tract relabeling) disrupts structural coherence in generated flows. To address these challenges, we propose Sat2Flow, a latent structure-aware diffusion-based framework that generates structurally coherent OD flows using solely satellite imagery as input. Our approach introduces a multi-kernel encoder to capture diverse regional interactions and employs a permutation-aware diffusion process that aligns latent representations across different regional orderings. Through a joint contrastive training objective that bridges satellite-derived features with OD patterns, combined with equivariant diffusion training that enforces structural consistency, Sat2Flow ensures topological robustness under arbitrary regional reindexing. Experimental results on real-world urban datasets demonstrate that Sat2Flow outperforms both physics-based and data-driven baselines in numerical accuracy while preserving empirical distributions and spatial structures under index permutations. Sat2Flow offers a globally scalable solution for OD flow generation in data-scarce urban environments, eliminating region-specific auxiliary data dependencies while maintaining structural invariance for robust mobility modeling.