Wenyuan Yu

Yuxing Xiang, Xue Li, Kun Qian et al.

With the widespread adoption of Large Language Models (LLMs), serving LLM inference requests has become an increasingly important task, attracting active research advancements. Practical workloads play an essential role in this process: they are critical for motivating and benchmarking serving techniques and systems. However, the existing understanding of real-world LLM serving workloads is limited due to the lack of a comprehensive workload characterization. Prior analyses remain insufficient in scale and scope, thus failing to fully capture intricate workload characteristics. In this paper, we fill the gap with an in-depth characterization of LLM serving workloads collected from our worldwide cloud inference serving service, covering not only language models but also emerging multimodal and reasoning models, and unveiling important new findings in each case. Moreover, based on our findings, we propose ServeGen, a principled framework for generating realistic LLM serving workloads by composing them on a per-client basis. A practical use case in production validates that ServeGen avoids 50% under-provisioning compared to naive workload generation, demonstrating ServeGen's advantage in performance benchmarking. ServeGen is available at https://github.com/alibaba/ServeGen.

Qian Tao, Zhen Wang, Wenyuan Yu et al.

In recent years, a plethora of spectral graph neural networks (GNN) methods have utilized polynomial basis with learnable coefficients to achieve top-tier performances on many node-level tasks. Although various kinds of polynomial bases have been explored, each such method adopts a fixed polynomial basis which might not be the optimal choice for the given graph. Besides, we identify the so-called over-passing issue of these methods and show that it is somewhat rooted in their less-principled regularization strategy and unnormalized basis. In this paper, we make the first attempts to address these two issues. Leveraging Jacobi polynomials, we design a novel spectral GNN, LON-GNN, with Learnable OrthoNormal bases and prove that regularizing coefficients becomes equivalent to regularizing the norm of learned filter function now. We conduct extensive experiments on diverse graph datasets to evaluate the fitting and generalization capability of LON-GNN, where the results imply its superiority.

Team Wan, Ang Wang, Baole Ai et al.

This report presents Wan, a comprehensive and open suite of video foundation models designed to push the boundaries of video generation. Built upon the mainstream diffusion transformer paradigm, Wan achieves significant advancements in generative capabilities through a series of innovations, including our novel VAE, scalable pre-training strategies, large-scale data curation, and automated evaluation metrics. These contributions collectively enhance the model's performance and versatility. Specifically, Wan is characterized by four key features: Leading Performance: The 14B model of Wan, trained on a vast dataset comprising billions of images and videos, demonstrates the scaling laws of video generation with respect to both data and model size. It consistently outperforms the existing open-source models as well as state-of-the-art commercial solutions across multiple internal and external benchmarks, demonstrating a clear and significant performance superiority. Comprehensiveness: Wan offers two capable models, i.e., 1.3B and 14B parameters, for efficiency and effectiveness respectively. It also covers multiple downstream applications, including image-to-video, instruction-guided video editing, and personal video generation, encompassing up to eight tasks. Consumer-Grade Efficiency: The 1.3B model demonstrates exceptional resource efficiency, requiring only 8.19 GB VRAM, making it compatible with a wide range of consumer-grade GPUs. Openness: We open-source the entire series of Wan, including source code and all models, with the goal of fostering the growth of the video generation community. This openness seeks to significantly expand the creative possibilities of video production in the industry and provide academia with high-quality video foundation models. All the code and models are available at https://github.com/Wan-Video/Wan2.1.

Chaojie Mao, Chen-Wei Xie, Chongyang Zhong et al.

We present Wan-Image, a unified visual generation system explicitly engineered to paradigm-shift image generation models from casual synthesizers into professional-grade productivity tools. While contemporary diffusion models excel at aesthetic generation, they frequently encounter critical bottlenecks in rigorous design workflows that demand absolute controllability, complex typography rendering, and strict identity preservation. To address these challenges, Wan-Image features a natively unified multi-modal architecture by synergizing the cognitive capabilities of large language models with the high-fidelity pixel synthesis of diffusion transformers, which seamlessly translates highly nuanced user intents into precise visual outputs. It is fundamentally powered by large-scale multi-modal data scaling, a systematic fine-grained annotation engine, and curated reinforcement learning data to surpass basic instruction following and unlock expert-level professional capabilities. These include ultra-long complex text rendering, hyper-diverse portrait generation, palette-guided generation, multi-subject identity preservation, coherent sequential visual generation, precise multi-modal interactive editing, native alpha-channel generation, and high-efficiency 4K synthesis. Across diverse human evaluations, Wan-Image exceeds Seedream 5.0 Lite and GPT Image 1.5 in overall performance, reaching parity with Nano Banana Pro in challenging tasks. Ultimately, Wan-Image revolutionizes visual content creation across e-commerce, entertainment, education, and personal productivity, redefining the boundaries of professional visual synthesis.

An Yang, Bowen Yu, Chengyuan Li et al.

We introduce Qwen2.5-1M, a series of models that extend the context length to 1 million tokens. Compared to the previous 128K version, the Qwen2.5-1M series have significantly enhanced long-context capabilities through long-context pre-training and post-training. Key techniques such as long data synthesis, progressive pre-training, and multi-stage supervised fine-tuning are employed to effectively enhance long-context performance while reducing training costs. To promote the use of long-context models among a broader user base, we present and open-source our inference framework. This framework includes a length extrapolation method that can expand the model context lengths by at least four times, or even more, without additional training. To reduce inference costs, we implement a sparse attention method along with chunked prefill optimization for deployment scenarios and a sparsity refinement method to improve precision. Additionally, we detail our optimizations in the inference engine, including kernel optimization, pipeline parallelism, and scheduling optimization, which significantly enhance overall inference performance. By leveraging our inference framework, the Qwen2.5-1M models achieve a remarkable 3x to 7x prefill speedup in scenarios with 1 million tokens of context. This framework provides an efficient and powerful solution for developing applications that require long-context processing using open-source models. The Qwen2.5-1M series currently includes the open-source models Qwen2.5-7B-Instruct-1M and Qwen2.5-14B-Instruct-1M, as well as the API-accessed model Qwen2.5-Turbo. Evaluations show that Qwen2.5-1M models have been greatly improved in long-context tasks without compromising performance in short-context scenarios. Specifically, the Qwen2.5-14B-Instruct-1M model significantly outperforms GPT-4o-mini in long-context tasks and supports contexts eight times longer.

Dingyan Zhang, Jinbo Han, Kaixi Zhang et al.

High-quality LLM request scheduling requires achieving two key objectives: whether the routed instance has KV$ to accelerate the request execution and whether the workload is balanced across instances. Achieving both objectives is challenging because pursuing one objective may compromise the other. Current approaches adopt various combinators (e.g., linear combinations) to compute a scheduling score combining indicators for the two objectives, which are complex in that they either require significant workload-specific hyperparameter tuning or model-hardware-aware simulator development, and could still lead to suboptimal performance. In this paper, we show that using a simple multiplication of two carefully chosen indicators-one for KV$-aware (new prefill tokens if routed to an instance) and one for load balancing-aware (current batch size of the instance)-as the scheduling score can simultaneously achieve both objectives well without any hyperparameter tuning. The key idea is that the multiplied score considers both objectives in a manner similar to a linear combination, with a nice property that the original hyperparameters are canceled out during comparison so we don't need tuning to find the best parameters. The two indicators are chosen based on our analysis of LLM characteristics, and our extensive experiments show that this simple approach can reduce TTFT by 92% and 52%, and TPOT by 21% and 20%, compared to vLLM-v1 and a production scheduler on real-world workloads covering chatbots, API calls, and coding agents. We also mathematically derive the conditions under which multiplication may fail, and find that such conditions are extremely rare in practice and can be detected (and mitigated) beforehand.

Yifan Peng, Jiafei Fan, Xingda Wei et al.

Learned top-K search is a promising approach for serving vector queries with both high accuracy and performance. However, current models trained for a specific K value fail to generalize to real-world multi-K queries: they suffer from accuracy degradation (for larger Ks) and performance loss (for smaller Ks). Training the model to generalize on different Ks requires orders of magnitude more preprocessing time and is not suitable for serving vector queries in the wild. We present OMEGA, a K-generalizable learned top-K search method that simultaneously achieves high accuracy, high performance, and low preprocessing cost for multi-K vector queries. The key idea is that a base model properly trained on K=1 with our trajectory-based features can be used to accurately predict larger Ks with a dynamic refinement procedure and smaller Ks with minimal performance loss. To make our refinements efficient, we further leverage the statistical properties of top-K searches to reduce excessive model invocations. Extensive evaluations on multiple public and production datasets show that, under the same preprocessing budgets, OMEGA achieves 6-33% lower average latency compared to state-of-the-art learned search methods, while all systems achieve the same recall target. With only 16-30% of the preprocessing time, OMEGA attains 1.01-1.28x of the optimal average latency of these baselines.

Lingkai Meng, Yu Shao, Long Yuan et al.

Usability evaluation is critical to the impact and adoption of open source software (OSS), yet traditional methods relying on human evaluators suffer from high costs and limited scalability. To address these limitations, we introduce OSS-UAgent, an automated, configurable, and interactive agent-based usability evaluation framework specifically designed for open source software. Our framework employs intelligent agents powered by large language models (LLMs) to simulate developers performing programming tasks across various experience levels (from Junior to Expert). By dynamically constructing platform-specific knowledge bases, OSS-UAgent ensures accurate and context-aware code generation. The generated code is automatically evaluated across multiple dimensions, including compliance, correctness, and readability, providing a comprehensive measure of the software's usability. Additionally, our demonstration showcases OSS-UAgent's practical application in evaluating graph analytics platforms, highlighting its effectiveness in automating usability evaluation.

Yunkai Lou, Shunyang Li, Longbin Lai et al.

Business Intelligence (BI) analysis is evolving towards Exploratory BI, an iterative, multi-round exploration paradigm where analysts progressively refine their understanding. However, traditional BI systems impose critical limits for Exploratory BI: heavy reliance on expert knowledge, high computational costs, static schemas, and lack of reusability. We present ExBI, a novel system that introduces the hypergraph data model with operators, including Source, Join, and View, to enable dynamic schema evolution and materialized view reuse. Using sampling-based algorithms with provable estimation guarantees, ExBI addresses the computational bottlenecks, while maintaining analytical accuracy. Experiments on LDBC datasets demonstrate that ExBI achieves significant speedups over existing systems: on average 16.21x (up to 146.25x) compared to Neo4j and 46.67x (up to 230.53x) compared to MySQL, while maintaining high accuracy with an average error rate of only 0.27% for COUNT, enabling efficient and accurate large-scale exploratory BI workflows.

Tao He, Xue Li, Zhibin Wang et al.

Training large-scale language models is increasingly critical in various domains, but it is hindered by frequent failures, leading to significant time and economic costs. Current failure recovery methods in cloud-based settings inadequately address the diverse and complex scenarios that arise, focusing narrowly on erasing downtime for individual tasks without considering the overall cost impact on a cluster. We introduce Unicron, a workload manager designed for efficient self-healing in large-scale language model training. Unicron optimizes the training process by minimizing failure-related costs across multiple concurrent tasks within a cluster. Its key features include in-band error detection for real-time error identification without extra overhead, a dynamic cost-aware plan generation mechanism for optimal reconfiguration, and an efficient transition strategy to reduce downtime during state changes. Deployed on a 128-GPU distributed cluster, Unicron demonstrates up to a 1.9x improvement in training efficiency over state-of-the-art methods, significantly reducing failure recovery costs and enhancing the reliability of large-scale language model training.

Qian Tao, Wenyuan Yu, Jingren Zhou

Large language models have shown exceptional capabilities in a wide range of tasks, such as text generation and video generation, among others. However, due to their massive parameter count, these models often require substantial storage space, imposing significant constraints on the machines deploying LLMs. To overcome this limitation, one research direction proposes to compress the models using integer replacements for floating-point numbers, in a process known as Quantization. Some recent studies suggest quantizing the key and value cache (KV Cache) of LLMs, and designing quantization techniques that treat the key and value matrices equivalently. This work delves deeper into the asymmetric structural roles of KV Cache, a phenomenon where the transformer's output loss is more sensitive to the quantization of key matrices. We conduct a systematic examination of the attention output error resulting from key and value quantization. The phenomenon inspires us to propose an asymmetric quantization strategy. Our approach allows for 1-bit quantization of the KV cache by implementing distinct configurations for key and value matrices. We carry out experiments across a variety of datasets, demonstrating that our proposed model allows for the quantization of up to 75% decoder layers with 1 bit, while simultaneously maintaining performance levels comparable to those of the models with floating parameters.

Jiahao Wang, Jinbo Han, Xingda Wei et al.

Serving large language models (LLMs) is important for cloud providers, and caching intermediate results (KV\$) after processing each request substantially improves serving throughput and latency. However, there is limited understanding of how LLM serving benefits from KV\$ caching, where system design decisions like cache eviction policies are highly workload-dependent. In this paper, we present the first systematic characterization of the KV\$ workload patterns from one of the leading LLM service providers. We draw observations that were not covered by previous studies focusing on synthetic workloads, including: KV\$ reuses are skewed across requests, where reuses between single-turn requests are equally important as multi-turn requests; the reuse time and probability are diverse considering all requests, but for a specific request category, the pattern tends to be predictable; and the overall cache size required for an ideal cache hit ratio is moderate. Based on the characterization, we further propose a workload-aware cache eviction policy that improves the serving performance under real-world traces, especially with limited cache capacity.

Haiquan Zhao, Chenhan Yuan, Fei Huang et al.

As large language models (LLMs) become more capable and widely used, ensuring the safety of their outputs is increasingly critical. Existing guardrail models, though useful in static evaluation settings, face two major limitations in real-world applications: (1) they typically output only binary "safe/unsafe" labels, which can be interpreted inconsistently across diverse safety policies, rendering them incapable of accommodating varying safety tolerances across domains; and (2) they require complete model outputs before performing safety checks, making them fundamentally incompatible with streaming LLM inference, thereby preventing timely intervention during generation and increasing exposure to harmful partial outputs. To address these challenges, we present Qwen3Guard, a series of multilingual safety guardrail models with two specialized variants: Generative Qwen3Guard, which casts safety classification as an instruction-following task to enable fine-grained tri-class judgments (safe, controversial, unsafe); and Stream Qwen3Guard, which introduces a token-level classification head for real-time safety monitoring during incremental text generation. Both variants are available in three sizes (0.6B, 4B, and 8B parameters) and support up to 119 languages and dialects, providing comprehensive, scalable, and low-latency safety moderation for global LLM deployments. Evaluated across English, Chinese, and multilingual benchmarks, Qwen3Guard achieves state-of-the-art performance in both prompt and response safety classification. All models are released under the Apache 2.0 license for public use.

Shenggan Cheng, Yuanxin Wei, Lansong Diao et al.

Leveraging the diffusion transformer (DiT) architecture, models like Sora, CogVideoX and Wan have achieved remarkable progress in text-to-video, image-to-video, and video editing tasks. Despite these advances, diffusion-based video generation remains computationally intensive, especially for high-resolution, long-duration videos. Prior work accelerates its inference by skipping computation, usually at the cost of severe quality degradation. In this paper, we propose SRDiffusion, a novel framework that leverages collaboration between large and small models to reduce inference cost. The large model handles high-noise steps to ensure semantic and motion fidelity (Sketching), while the smaller model refines visual details in low-noise steps (Rendering). Experimental results demonstrate that our method outperforms existing approaches, over 3$\times$ speedup for Wan with nearly no quality loss for VBench, and 2$\times$ speedup for CogVideoX. Our method is introduced as a new direction orthogonal to existing acceleration strategies, offering a practical solution for scalable video generation.

Yinghan Li, Yifei Li, Jiejing Zhang et al.

It has long been a problem to arrange and execute irregular workloads on massively parallel devices. We propose a general framework for statically batching irregular workloads into a single kernel with a runtime task mapping mechanism on GPUs. We further apply this framework to Mixture-of-Experts (MoE) model inference and implement an optimized and efficient CUDA kernel. Our MoE kernel achieves up to 91% of the peak Tensor Core throughput on NVIDIA H800 GPU and 95% on NVIDIA H20 GPU.

Yuanxin Wei, Lansong Diao, Bujiao Chen et al.

Leveraging the Transformer architecture and the diffusion process, video DiT models have emerged as a dominant approach for high-quality video generation. However, their multi-step iterative denoising process incurs high computational cost and inference latency. Caching, a widely adopted optimization method in DiT models, leverages the redundancy in the diffusion process to skip computations in different granularities (e.g., step, cfg, block). Nevertheless, existing caching methods are limited to single-granularity strategies, struggling to balance generation quality and inference speed in a flexible manner. In this work, we propose MixCache, a training-free caching-based framework for efficient video DiT inference. It first distinguishes the interference and boundary between different caching strategies, and then introduces a context-aware cache triggering strategy to determine when caching should be enabled, along with an adaptive hybrid cache decision strategy for dynamically selecting the optimal caching granularity. Extensive experiments on diverse models demonstrate that, MixCache can significantly accelerate video generation (e.g., 1.94$\times$ speedup on Wan 14B, 1.97$\times$ speedup on HunyuanVideo) while delivering both superior generation quality and inference efficiency compared to baseline methods.

Qian Tao, Xiyuan Wang, Muhan Zhang et al.

Graph neural networks (GNNs) have become a prevalent framework for graph tasks. Many recent studies have proposed the use of graph convolution methods over the numerous subgraphs of each graph, a concept known as subgraph graph neural networks (subgraph GNNs), to enhance GNNs' ability to distinguish non-isomorphic graphs. To maximize the expressiveness, subgraph GNNs often require each subgraph to have equal size to the original graph. Despite their impressive performance, subgraph GNNs face challenges due to the vast number and large size of subgraphs which lead to a surge in training data, resulting in both storage and computational inefficiencies. In response to this problem, this paper introduces Ego-Nets-Fit-All (ENFA), a model that uniformly takes the smaller ego nets as subgraphs, thereby providing greater storage and computational efficiency, while at the same time guarantees identical outputs to the original subgraph GNNs even taking the whole graph as subgraphs. The key is to identify and eliminate the redundant computation among subgraphs. For example, a node $v_i$ may appear in multiple subgraphs but is far away from all of their centers (the unsymmetric part between subgraphs). Therefore, its first few rounds of message passing within each subgraph can be computed once in the original graph instead of being computed multiple times within each subgraph. Such strategy enables our ENFA to accelerate subgraph GNNs in an exact way, unlike previous sampling approaches that often lose the performance. Extensive experiments across various datasets reveal that compared with the conventional subgraph GNNs, ENFA can reduce storage space by 29.0% to 84.5% and improve training efficiency by up to 1.66x.