Yilin Zhou

GLM-5 Team, Aohan Zeng, Xin Lv et al. · tsinghua

We present GLM-5, a next-generation foundation model designed to transition the paradigm of vibe coding to agentic engineering. Building upon the agentic, reasoning, and coding (ARC) capabilities of its predecessor, GLM-5 adopts DSA to significantly reduce training and inference costs while maintaining long-context fidelity. To advance model alignment and autonomy, we implement a new asynchronous reinforcement learning infrastructure that drastically improves post-training efficiency by decoupling generation from training. Furthermore, we propose novel asynchronous agent RL algorithms that further improve RL quality, enabling the model to learn from complex, long-horizon interactions more effectively. Through these innovations, GLM-5 achieves state-of-the-art performance on major open benchmarks. Most critically, GLM-5 demonstrates unprecedented capability in real-world coding tasks, surpassing previous baselines in handling end-to-end software engineering challenges. Code, models, and more information are available at https://github.com/zai-org/GLM-5.

Wenhe Jia, Lu Yang, Zilong Jia et al.

Since first proposed, Video Instance Segmentation(VIS) task has attracted vast researchers' focus on architecture modeling to boost performance. Though great advances achieved in online and offline paradigms, there are still insufficient means to identify model errors and distinguish discrepancies between methods, as well approaches that correctly reflect models' performance in recognizing object instances of various temporal lengths remain barely available. More importantly, as the fundamental model abilities demanded by the task, spatial segmentation and temporal association are still understudied in both evaluation and interaction mechanisms. In this paper, we introduce TIVE, a Toolbox for Identifying Video instance segmentation Errors. By directly operating output prediction files, TIVE defines isolated error types and weights each type's damage to mAP, for the purpose of distinguishing model characters. By decomposing localization quality in spatial-temporal dimensions, model's potential drawbacks on spatial segmentation and temporal association can be revealed. TIVE can also report mAP over instance temporal length for real applications. We conduct extensive experiments by the toolbox to further illustrate how spatial segmentation and temporal association affect each other. We expect the analysis of TIVE can give the researchers more insights, guiding the community to promote more meaningful explorations for video instance segmentation. The proposed toolbox is available at https://github.com/wenhe-jia/TIVE.

GLM-4. 5 Team, Aohan Zeng, Xin Lv et al.

We present GLM-4.5, an open-source Mixture-of-Experts (MoE) large language model with 355B total parameters and 32B activated parameters, featuring a hybrid reasoning method that supports both thinking and direct response modes. Through multi-stage training on 23T tokens and comprehensive post-training with expert model iteration and reinforcement learning, GLM-4.5 achieves strong performance across agentic, reasoning, and coding (ARC) tasks, scoring 70.1% on TAU-Bench, 91.0% on AIME 24, and 64.2% on SWE-bench Verified. With much fewer parameters than several competitors, GLM-4.5 ranks 3rd overall among all evaluated models and 2nd on agentic benchmarks. We release both GLM-4.5 (355B parameters) and a compact version, GLM-4.5-Air (106B parameters), to advance research in reasoning and agentic AI systems. Code, models, and more information are available at https://github.com/zai-org/GLM-4.5.

Siyi Chen, Tianhan Luo, Shijian Wu et al.

Open-source libraries are widely used in modern software development, introducing significant security vulnerabilities. While static analysis tools can identify potential vulnerabilities at scale, they often generate overwhelming reports with high false positive rates. Automated Exploit Generation (AEG) emerges as a promising solution to confirm vulnerability authenticity by generating an exploit. However, traditional AEG approaches based on fuzzing or symbolic execution face path coverage and constraint-solving problems. Although LLMs show great potential for AEG, how to effectively leverage them to comprehend vulnerabilities and generate corresponding exploits is still an open question. To address these challenges, we propose Vulnsage, a multi-agent framework for AEG. Vulnsage simulates human security researchers' workflows by decomposing the complex AEG process into multiple specialized sub-agents: Code Analyzer Agent, Code Generation Agent, Validation Agent, and a set of Reflection Agents, orchestrated by a central supervisor through iterative cycles. Given a target program, the Code Analyzer Agent performs static analysis to identify potential vulnerabilities and collects relevant information for each one. The Code Generation Agent then utilizes an LLM to generate candidate exploits. The Validation Agent and Reflection Agents form a feedback-driven self-refinement loop that uses execution traces and runtime error analysis to either improve the exploit iteratively or reason about the false positive alert. Experimental evaluation demonstrates that Vulnsage succeeds in generating 34.64\% more exploits than state-of-the-art tools such as \explodejs. Furthermore, Vulnsage has successfully discovered and verified 146 zero-day vulnerabilities in real-world scenarios, demonstrating its practical effectiveness for assisting security assessment in software supply chains.

Wenhe Jia, Yilin Zhou, Xuhan Zhu et al.

Dense pose estimation is a dense 3D prediction task for instance-level human analysis, aiming to map human pixels from an RGB image to a 3D surface of the human body. Due to a large amount of surface point regression, the training process appears to be easy to collapse compared to other region-based human instance analyzing tasks. By analyzing the loss formulation of the existing dense pose estimation model, we introduce a novel point regression loss function, named Dense Points} loss to stable the training progress, and a new balanced loss weighting strategy to handle the multi-task losses. With the above novelties, we propose a brand new architecture, named UV R-CNN. Without auxiliary supervision and external knowledge from other tasks, UV R-CNN can handle many complicated issues in dense pose model training progress, achieving 65.0% $AP_{gps}$ and 66.1% $AP_{gpsm}$ on the DensePose-COCO validation subset with ResNet-50-FPN feature extractor, competitive among the state-of-the-art dense human pose estimation methods.

Andrew Zhuoer Feng, Cunxiang Wang, Yu Luo et al.

Evaluating the writing capabilities of large language models (LLMs) remains a significant challenge due to the multidimensional nature of writing skills and the limitations of existing metrics. LLM's performance in thousand-words level and open-ended writing is inadequately assessed by traditional reference-based metrics or modern LLM-as-a-judge methods. We propose Tree-of-Writing (ToW), to resolve the implicit inconsistency often found when LLM-as-a-judge aggregates all sub-features in text evaluation. ToW incorporates a tree-structured workflow by explicitly modeling the aggregation weights of sub-features. We also present HowToBench, a large-scale Chinese writing benchmark encompassing 12 genres and 1302 instructions across three task categories: contextual completion, outline-guided writing, and open-ended generation. ToW successfully mitigates the biases, achieving a 0.93 Pearson correlation with human judgments. Furthermore, we detect that both overlap-based text generation metrics and popular LLM-as-a-judge practices are vulnerable to textual disturbances, while ToW is robust to them. We also uncover a negative correlation between input length and content-related scores in the Guide task, showcasing that it cannot be simply improved by input-side information piling.

YiLin Zhou, Lili Wei, Zheming Xu et al.

Few-shot 3D point cloud semantic segmentation (FS-3DSeg) aims to segment novel classes with only a few labeled samples. However, existing metric-based prototype learning methods generate prototypes solely from the support set, without considering their relevance to query data. This often results in prototype bias, where prototypes overfit support-specific characteristics and fail to generalize to the query distribution, especially in the presence of distribution shifts, which leads to degraded segmentation performance. To address this issue, we propose a novel Query-aware Hub Prototype (QHP) learning method that explicitly models semantic correlations between support and query sets. Specifically, we propose a Hub Prototype Generation (HPG) module that constructs a bipartite graph connecting query and support points, identifies frequently linked support hubs, and generates query-relevant prototypes that better capture cross-set semantics. To further mitigate the influence of bad hubs and ambiguous prototypes near class boundaries, we introduce a Prototype Distribution Optimization (PDO) module, which employs a purity-reweighted contrastive loss to refine prototype representations by pulling bad hubs and outlier prototypes closer to their corresponding class centers. Extensive experiments on S3DIS and ScanNet demonstrate that QHP achieves substantial performance gains over state-of-the-art methods, effectively narrowing the semantic gap between prototypes and query sets in FS-3DSeg.

Mingxian Li, Hao Sun, Yingtie Lei et al.

Document images are often degraded by various stains, significantly impacting their readability and hindering downstream applications such as document digitization and analysis. The absence of a comprehensive stained document dataset has limited the effectiveness of existing document enhancement methods in removing stains while preserving fine-grained details. To address this challenge, we construct StainDoc, the first large-scale, high-resolution ($2145\times2245$) dataset specifically designed for document stain removal. StainDoc comprises over 5,000 pairs of stained and clean document images across multiple scenes. This dataset encompasses a diverse range of stain types, severities, and document backgrounds, facilitating robust training and evaluation of document stain removal algorithms. Furthermore, we propose StainRestorer, a Transformer-based document stain removal approach. StainRestorer employs a memory-augmented Transformer architecture that captures hierarchical stain representations at part, instance, and semantic levels via the DocMemory module. The Stain Removal Transformer (SRTransformer) leverages these feature representations through a dual attention mechanism: an enhanced spatial attention with an expanded receptive field, and a channel attention captures channel-wise feature importance. This combination enables precise stain removal while preserving document content integrity. Extensive experiments demonstrate StainRestorer's superior performance over state-of-the-art methods on the StainDoc dataset and its variants StainDoc\_Mark and StainDoc\_Seal, establishing a new benchmark for document stain removal. Our work highlights the potential of memory-augmented Transformers for this task and contributes a valuable dataset to advance future research.