TraceSIR: A Multi-Agent Framework for Structured Analysis and Reporting of Agentic Execution Traces

Agentic systems augment large language models with external tools and iterative decision making, enabling complex tasks such as deep research, function calling, and coding. However, their long and intricate execution traces make failure diagnosis and root cause analysis extremely challenging. Manual inspection does not scale, while directly applying LLMs to raw traces is hindered by input length limits and unreliable reasoning. Focusing solely on final task outcomes further discards critical behavioral information required for accurate issue localization. To address these issues, we propose TraceSIR, a multi-agent framework for structured analysis and reporting of agentic execution traces. TraceSIR coordinates three specialized agents: (1) StructureAgent, which introduces a novel abstraction format, TraceFormat, to compress execution traces while preserving essential behavioral information; (2) InsightAgent, which performs fine-grained diagnosis including issue localization, root cause analysis, and optimization suggestions; (3) ReportAgent, which aggregates insights across task instances and generates comprehensive analysis reports. To evaluate TraceSIR, we construct TraceBench, covering three real-world agentic scenarios, and introduce ReportEval, an evaluation protocol for assessing the quality and usability of analysis reports aligned with industry needs. Experiments show that TraceSIR consistently produces coherent, informative, and actionable reports, significantly outperforming existing approaches across all evaluation dimensions. Our project and video are publicly available at https://github.com/SHU-XUN/TraceSIR.

RAVEL: Reasoning Agents for Validating and Evaluating LLM Text Synthesis

Large Language Models have evolved from single-round generators into long-horizon agents, capable of complex text synthesis scenarios. However, current evaluation frameworks lack the ability to assess the actual synthesis operations, such as outlining, drafting, and editing. Consequently, they fail to evaluate the actual and detailed capabilities of LLMs. To bridge this gap, we introduce RAVEL, an agentic framework that enables the LLM testers to autonomously plan and execute typical synthesis operations, including outlining, drafting, reviewing, and refining. Complementing this framework, we present C3EBench, a comprehensive benchmark comprising 1,258 samples derived from professional human writings. We utilize a "reverse-engineering" pipeline to isolate specific capabilities across four tasks: Cloze, Edit, Expand, and End-to-End. Through our analysis of 14 LLMs, we uncover that most LLMs struggle with tasks that demand contextual understanding under limited or under-specified instructions. By augmenting RAVEL with SOTA LLMs as operators, we find that such agentic text synthesis is dominated by the LLM's reasoning capability rather than raw generative capacity. Furthermore, we find that a strong reasoner can guide a weaker generator to yield higher-quality results, whereas the inverse does not hold. Our code and data are available at this link: https://github.com/ZhuoerFeng/RAVEL-Reasoning-Agents-Text-Eval.

GeCo-SRT: Geometry-aware Continual Adaptation for Robotic Cross-Task Sim-to-Real Transfer

Bridging the sim-to-real gap is important for applying low-cost simulation data to real-world robotic systems. However, previous methods are severely limited by treating each transfer as an isolated endeavor, demanding repeated, costly tuning and wasting prior transfer experience. To move beyond isolated sim-to-real, we build a continual cross-task sim-to-real transfer paradigm centered on knowledge accumulation across iterative transfers, thereby enabling effective and efficient adaptation to novel tasks. Thus, we propose GeCo-SRT, a geometry-aware continual adaptation method. It utilizes domain-invariant and task-invariant knowledge from local geometric features as a transferable foundation to accelerate adaptation during subsequent sim-to-real transfers. This method starts with a geometry-aware mixture-of-experts module, which dynamically activates experts to specialize in distinct geometric knowledge to bridge observation sim-to-real gap. Further, the geometry-expert-guided prioritized experience replay module preferentially samples from underutilized experts, refreshing specialized knowledge to combat forgetting and maintain robust cross-task performance. Leveraging knowledge accumulated during iterative transfer, GeCo-SRT method not only achieves 52% average performance improvement over the baseline, but also demonstrates significant data efficiency for new task adaptation with only 1/6 data. We hope this work inspires approaches for efficient, low-cost cross-task sim-to-real transfer.

GLM-5: from Vibe Coding to Agentic Engineering

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.

Beyond Literal Mapping: Benchmarking and Improving Non-Literal Translation Evaluation

Large Language Models (LLMs) have significantly advanced Machine Translation (MT), applying them to linguistically complex domains-such as Social Network Services, literature etc. In these scenarios, translations often require handling non-literal expressions, leading to the inaccuracy of MT metrics. To systematically investigate the reliability of MT metrics, we first curate a meta-evaluation dataset focused on non-literal translations, namely MENT. MENT encompasses four non-literal translation domains and features source sentences paired with translations from diverse MT systems, with 7,530 human-annotated scores on translation quality. Experimental results reveal the inaccuracies of traditional MT metrics and the limitations of LLM-as-a-Judge, particularly the knowledge cutoff and score inconsistency problem. To mitigate these limitations, we propose RATE, a novel agentic translation evaluation framework, centered by a reflective Core Agent that dynamically invokes specialized sub-agents. Experimental results indicate the efficacy of RATE, achieving an improvement of at least 3.2 meta score compared with current metrics. Further experiments demonstrate the robustness of RATE to general-domain MT evaluation. Code and dataset are available at: https://github.com/BITHLP/RATE.

Every Step Evolves: Scaling Reinforcement Learning for Trillion-Scale Thinking Model

We present Ring-1T, the first open-source, state-of-the-art thinking model with a trillion-scale parameter. It features 1 trillion total parameters and activates approximately 50 billion per token. Training such models at a trillion-parameter scale introduces unprecedented challenges, including train-inference misalignment, inefficiencies in rollout processing, and bottlenecks in the RL system. To address these, we pioneer three interconnected innovations: (1) IcePop stabilizes RL training via token-level discrepancy masking and clipping, resolving instability from training-inference mismatches; (2) C3PO++ improves resource utilization for long rollouts under a token budget by dynamically partitioning them, thereby obtaining high time efficiency; and (3) ASystem, a high-performance RL framework designed to overcome the systemic bottlenecks that impede trillion-parameter model training. Ring-1T delivers breakthrough results across critical benchmarks: 93.4 on AIME-2025, 86.72 on HMMT-2025, 2088 on CodeForces, and 55.94 on ARC-AGI-v1. Notably, it attains a silver medal-level result on the IMO-2025, underscoring its exceptional reasoning capabilities. By releasing the complete 1T parameter MoE model to the community, we provide the research community with direct access to cutting-edge reasoning capabilities. This contribution marks a significant milestone in democratizing large-scale reasoning intelligence and establishes a new baseline for open-source model performance.

GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models

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.

Editable-DeepSC: Reliable Cross-Modal Semantic Communications for Facial Editing

Real-time computer vision (CV) plays a crucial role in various real-world applications, whose performance is highly dependent on communication networks. Nonetheless, the data-oriented characteristics of conventional communications often do not align with the special needs of real-time CV tasks. To alleviate this issue, the recently emerged semantic communications only transmit task-related semantic information and exhibit a promising landscape to address this problem. However, the communication challenges associated with Semantic Facial Editing, one of the most important real-time CV applications on social media, still remain largely unexplored. In this paper, we fill this gap by proposing Editable-DeepSC, a novel cross-modal semantic communication approach for facial editing. Firstly, we theoretically discuss different transmission schemes that separately handle communications and editings, and emphasize the necessity of Joint Editing-Channel Coding (JECC) via iterative attributes matching, which integrates editings into the communication chain to preserve more semantic mutual information. To compactly represent the high-dimensional data, we leverage inversion methods via pre-trained StyleGAN priors for semantic coding. To tackle the dynamic channel noise conditions, we propose SNR-aware channel coding via model fine-tuning. Extensive experiments indicate that Editable-DeepSC can achieve superior editings while significantly saving the transmission bandwidth, even under high-resolution and out-of-distribution (OOD) settings.

MIBench: A Comprehensive Benchmark for Model Inversion Attack and Defense

Model Inversion (MI) attacks aim at leveraging the output information of target models to reconstruct privacy-sensitive training data, raising widespread concerns on privacy threats of Deep Neural Networks (DNNs). Unfortunately, in tandem with the rapid evolution of MI attacks, the lack of a comprehensive, aligned, and reliable benchmark has emerged as a formidable challenge. This deficiency leads to inadequate comparisons between different attack methods and inconsistent experimental setups. In this paper, we introduce the first practical benchmark for model inversion attacks and defenses to address this critical gap, which is named \textit{MIBench}. This benchmark serves as an extensible and reproducible modular-based toolbox and currently integrates a total of 16 state-of-the-art attack and defense methods. Moreover, we furnish a suite of assessment tools encompassing 9 commonly used evaluation protocols to facilitate standardized and fair evaluation and analysis. Capitalizing on this foundation, we conduct extensive experiments from multiple perspectives to holistically compare and analyze the performance of various methods across different scenarios, which overcomes the misalignment issues and discrepancy prevalent in previous works. Based on the collected attack methods and defense strategies, we analyze the impact of target resolution, defense robustness, model predictive power, model architectures, transferability and loss function. Our hope is that this \textit{MIBench} could provide a unified, practical and extensible toolbox and is widely utilized by researchers in the field to rigorously test and compare their novel methods, ensuring equitable evaluations and thereby propelling further advancements in the future development.

One Perturbation is Enough: On Generating Universal Adversarial Perturbations against Vision-Language Pre-training Models

Vision-Language Pre-training (VLP) models trained on large-scale image-text pairs have demonstrated unprecedented capability in many practical applications. However, previous studies have revealed that VLP models are vulnerable to adversarial samples crafted by a malicious adversary. While existing attacks have achieved great success in improving attack effect and transferability, they all focus on instance-specific attacks that generate perturbations for each input sample. In this paper, we show that VLP models can be vulnerable to a new class of universal adversarial perturbation (UAP) for all input samples. Although initially transplanting existing UAP algorithms to perform attacks showed effectiveness in attacking discriminative models, the results were unsatisfactory when applied to VLP models. To this end, we revisit the multimodal alignments in VLP model training and propose the Contrastive-training Perturbation Generator with Cross-modal conditions (C-PGC). Specifically, we first design a generator that incorporates cross-modal information as conditioning input to guide the training. To further exploit cross-modal interactions, we propose to formulate the training objective as a multimodal contrastive learning paradigm based on our constructed positive and negative image-text pairs. By training the conditional generator with the designed loss, we successfully force the adversarial samples to move away from its original area in the VLP model's feature space, and thus essentially enhance the attacks. Extensive experiments show that our method achieves remarkable attack performance across various VLP models and Vision-and-Language (V+L) tasks. Moreover, C-PGC exhibits outstanding black-box transferability and achieves impressive results in fooling prevalent large VLP models including LLaVA and Qwen-VL.