GrandCode: Achieving Grandmaster Level in Competitive Programming via Agentic Reinforcement Learning
DeepReinforce
GrandCode：通过智能体强化学习达到竞赛编程特级大师水平
Competitive programming remains one of the last few human strongholds in coding against AI. The best AI system to date still underperforms the best humans competitive programming: the most recent best result, Google's Gemini~3 Deep Think, attained 8th place even not being evaluated under live competition conditions. In this work, we introduce GrandCode, a multi-agent RL system designed for competitive programming. The capability of GrandCode is attributed to two key factors: (1) It orchestrates a variety of agentic modules (hypothesis proposal, solver, test generator, summarization, etc) and jointly improves them through post-training and online test-time RL; (2) We introduce Agentic GRPO specifically designed for multi-stage agent rollouts with delayed rewards and the severe off-policy drift that is prevalent in agentic RL. GrandCode is the first AI system that consistently beats all human participants in live contests of competitive programming: in the most recent three Codeforces live competitions, i.e., Round~1087 (Mar 21, 2026), Round~1088 (Mar 28, 2026), and Round~1089 (Mar 29, 2026), GrandCode placed first in all of them, beating all human participants, including legendary grandmasters. GrandCode shows that AI systems have reached a point where they surpass the strongest human programmers on the most competitive coding tasks.
竞赛编程仍是人类在编程领域对抗 AI 的少数堡垒之一。迄今最强的 AI 系统依然不及顶尖人类选手：Google Gemini 3 Deep Think 的最新最佳成绩仅为第 8 名，而且并非在实时比赛条件下评测。本文提出 GrandCode，一个面向竞赛编程的多智能体强化学习系统。GrandCode 的能力源自两个关键因素：其一，它编排假设提出器、求解器、测试生成器、总结器等多种智能体模块，并通过后训练和在线测试时强化学习联合提升这些模块；其二，我们提出 Agentic GRPO，专门处理多阶段智能体轨迹中的延迟奖励和智能体强化学习中常见的严重离策略漂移。GrandCode 是首个能在实时竞赛编程比赛中持续击败全部人类参赛者的 AI 系统：在最近三场 Codeforces 实时比赛 Round 1087、Round 1088 和 Round 1089 中，它均获得第一名，战胜了包括传奇特级大师在内的所有人类选手。GrandCode 表明，AI 系统已能在竞争最激烈的编程任务上超越最强人类程序员。
Adam's Law: Textual Frequency Law on Large Language Models
FaceMind
亚当定律：大语言模型的文本频率定律
While textual frequency has been validated as relevant to human cognition in reading speed, its relatedness to Large Language Models (LLMs) is seldom studied. We propose a novel research direction in terms of textual data frequency, which is an understudied topic, to the best of our knowledge. Our framework is composed of three units. First, this paper proposes Textual Frequency Law (TFL), which indicates that frequent textual data should be preferred for LLMs for both prompting and fine-tuning. Since many LLMs are closed-source in their training data, we propose using online resources to estimate the sentence-level frequency. We then utilize an input paraphraser to paraphrase the input into a more frequent textual expression. Next, we propose Textual Frequency Distillation (TFD) by querying LLMs to conduct story completion by further extending the sentences in the datasets, and the resulting corpora are used to adjust the initial estimation. Finally, we propose Curriculum Textual Frequency Training (CTFT) that fine-tunes LLMs in an increasing order of sentence-level frequency. Experiments are conducted on our curated dataset Textual Frequency Paired Dataset (TFPD) on math reasoning, machine translation, commonsense reasoning and agentic tool calling. Results show the effectiveness of our framework.
文本频率已被证实与人类阅读速度等认知活动有关，但它与大语言模型之间的关系却鲜有研究。据我们所知，本文从文本数据频率出发提出了一个尚未得到充分探索的新研究方向。我们的框架由三个部分组成。首先，本文提出文本频率定律（TFL），指出无论是提示还是微调，大语言模型都应优先使用高频文本数据。由于许多大语言模型并未公开训练数据，我们提出利用在线资源估计句子级频率，再通过输入改写器将输入改写为更常见的文本表达。其次，我们提出文本频率蒸馏（TFD）：让大语言模型通过续写数据集中的句子完成故事，并使用所得语料修正初始频率估计。最后，我们提出课程式文本频率训练（CTFT），按照句子级频率递增的顺序微调大语言模型。我们在构建的文本频率配对数据集（TFPD）上开展实验，覆盖数学推理、机器翻译、常识推理和智能体工具调用。结果验证了该框架的有效性。
DataFlex: A Unified Framework for Data-Centric Dynamic Training of Large Language Models
北京大学
DataFlex：面向大语言模型数据中心动态训练的统一框架
Data-centric training has emerged as a promising direction for improving large language models (LLMs) by optimizing not only model parameters but also the selection, composition, and weighting of training data during optimization. However, existing approaches to data selection, data mixture optimization, and data reweighting are often developed in isolated codebases with inconsistent interfaces, hindering reproducibility, fair comparison, and practical integration. In this paper, we present DataFlex, a unified data-centric dynamic training framework built upon LLaMA-Factory. DataFlex supports three major paradigms of dynamic data optimization: sample selection, domain mixture adjustment, and sample reweighting, while remaining fully compatible with the original training workflow. It provides extensible trainer abstractions and modular components, enabling a drop-in replacement for standard LLM training, and unifies key model-dependent operations such as embedding extraction, inference, and gradient computation, with support for large-scale settings including DeepSpeed ZeRO-3. We conduct comprehensive experiments across multiple data-centric methods. Dynamic data selection consistently outperforms static full-data training on MMLU across both Mistral-7B and Llama-3.2-3B. For data mixture, DoReMi and ODM improve both MMLU accuracy and corpus-level perplexity over default proportions when pretraining Qwen2.5-1.5B on SlimPajama at 6B and 30B token scales. DataFlex also achieves consistent runtime improvements over original implementations. These results demonstrate that DataFlex provides an effective, efficient, and reproducible infrastructure for data-centric dynamic training of LLMs.
数据中心训练已成为改进大语言模型的一条重要路径：在优化过程中，它不仅优化模型参数，也优化训练数据的选择、组成与权重。然而，现有的数据选择、数据混合优化和数据重加权方法往往分散在接口不一致的独立代码库中，妨碍了结果复现、公平比较和实际集成。本文提出 DataFlex，一个基于 LLaMA-Factory 构建的统一数据中心动态训练框架。DataFlex 支持样本选择、领域混合调整和样本重加权三种主要的动态数据优化范式，同时完全兼容原有训练流程。它提供可扩展的训练器抽象和模块化组件，可直接替代标准大语言模型训练流程；同时统一嵌入提取、推理和梯度计算等依赖模型的关键操作，并支持 DeepSpeed ZeRO-3 等大规模训练设置。我们对多种数据中心方法进行了全面实验。在 Mistral-7B 和 Llama-3.2-3B 上，动态数据选择在 MMLU 上始终优于静态全量数据训练。在数据混合方面，以 60 亿和 300 亿 token 在 SlimPajama 上预训练 Qwen2.5-1.5B 时，DoReMi 和 ODM 相比默认比例均提升了 MMLU 准确率并降低了语料级困惑度。DataFlex 相比原始实现也取得了稳定的运行时间改进。结果表明，DataFlex 为大语言模型的数据中心动态训练提供了有效、高效且可复现的基础设施。
FIPO: Eliciting Deep Reasoning with Future-KL Influenced Policy Optimization
Qwen
FIPO：通过未来 KL 影响的策略优化激发深度推理
We present Future-KL Influenced Policy Optimization (FIPO), a reinforcement learning algorithm designed to overcome reasoning bottlenecks in large language models. While GRPO style training scales effectively, it typically relies on outcome-based rewards (ORM) that distribute a global advantage uniformly across every token in a trajectory. We argue that this coarse-grained credit assignment imposes a performance ceiling by failing to distinguish critical logical pivots from trivial tokens. FIPO addresses this by incorporating discounted future-KL divergence into the policy update, creating a dense advantage formulation that re-weights tokens based on their influence on subsequent trajectory behavior. Empirically, FIPO enables models to break through the length stagnation seen in standard baselines. Evaluated on Qwen2.5-32B, FIPO extends the average chain-of-thought length from roughly 4,000 to over 10,000 tokens and increases AIME 2024 Pass@1 accuracy from 50.0% to a peak of 58.0% (converging at approximately 56.0%). This outperforms both DeepSeek-R1-Zero-Math-32B (around 47.0%) and o1-mini (approximately 56.0%). Our results suggest that establishing dense advantage formulations is a vital path for evolving ORM-based algorithms to unlock the full reasoning potential of base models. We open-source our training system, built on the verl framework.
我们提出未来 KL 影响的策略优化（FIPO），一种旨在突破大语言模型推理瓶颈的强化学习算法。GRPO 风格训练虽然具有良好的扩展性，但通常依赖结果奖励模型（ORM），将同一个全局优势均匀分配给轨迹中的每个 token。我们认为，这种粗粒度信用分配无法区分关键逻辑转折与无关紧要的 token，因而限制了性能上限。FIPO 将折扣未来 KL 散度纳入策略更新，构造一种稠密优势形式，根据 token 对后续轨迹行为的影响重新赋权。实验表明，FIPO 能使模型突破标准基线中的长度停滞。在 Qwen2.5-32B 上，FIPO 将平均思维链长度从约 4,000 个 token 延长到超过 10,000 个，并将 AIME 2024 Pass@1 准确率从 50.0% 提升至最高 58.0%，最终收敛在约 56.0%。这一结果超过 DeepSeek-R1-Zero-Math-32B 的约 47.0% 和 o1-mini 的约 56.0%。结果表明，建立稠密优势形式是推动基于 ORM 的算法演进、释放基础模型完整推理潜力的重要方向。我们还开源了基于 verl 框架构建的训练系统。
CARLA-Air: Fly Drones Inside a CARLA World -- A Unified Infrastructure for Air-Ground Embodied Intelligence
CARLA-Air：在 CARLA 世界中飞行无人机——面向空地具身智能的统一基础设施
The convergence of low-altitude economies, embodied intelligence, and air-ground cooperative systems creates growing demand for simulation infrastructure capable of jointly modeling aerial and ground agents within a single physically coherent environment. Existing open-source platforms remain domain-segregated: driving simulators lack aerial dynamics, while multirotor simulators lack realistic ground scenes. Bridge-based co-simulation introduces synchronization overhead and cannot guarantee strict spatial-temporal consistency. We present CARLA-Air, an open-source infrastructure that unifies high-fidelity urban driving and physics-accurate multirotor flight within a single Unreal Engine process. The platform preserves both CARLA and AirSim native Python APIs and ROS 2 interfaces, enabling zero-modification code reuse. Within a shared physics tick and rendering pipeline, CARLA-Air delivers photorealistic environments with rule-compliant traffic, socially-aware pedestrians, and aerodynamically consistent UAV dynamics, synchronously capturing up to 18 sensor modalities across all platforms at each tick. The platform supports representative air-ground embodied intelligence workloads spanning cooperation, embodied navigation and vision-language action, multi-modal perception and dataset construction, and reinforcement-learning-based policy training. An extensible asset pipeline allows integration of custom robot platforms into the shared world. By inheriting AirSim's aerial capabilities -- whose upstream development has been archived -- CARLA-Air ensures this widely adopted flight stack continues to evolve within a modern infrastructure. Released with prebuilt binaries and full source: https://github.com/louiszengCN/CarlaAir
低空经济、具身智能和空地协同系统的融合，使人们日益需要能够在同一个物理一致环境中联合建模空中与地面智能体的仿真基础设施。现有开源平台仍彼此割裂：自动驾驶仿真器缺乏飞行动力学，多旋翼仿真器则缺少真实地面场景；基于桥接的协同仿真不仅会引入同步开销，也无法保证严格的时空一致性。我们提出 CARLA-Air，一个在单个 Unreal Engine 进程中统一高保真城市驾驶与物理精确多旋翼飞行的开源基础设施。该平台保留 CARLA 和 AirSim 原生 Python API 与 ROS 2 接口，使已有代码无需修改即可复用。在共享的物理时钟与渲染管线中，CARLA-Air 提供具有规则化交通、具备社会感知能力的行人和空气动力学一致无人机动态的照片级真实环境，并能在每个时钟周期同步采集所有平台上最多 18 种传感器模态。平台支持空地协作、具身导航与视觉语言行动、多模态感知与数据集构建，以及基于强化学习的策略训练等代表性空地具身智能工作负载。可扩展资产管线还允许将自定义机器人平台集成到共享世界中。CARLA-Air 继承了上游开发已经归档的 AirSim 飞行能力，使这一广泛采用的飞行技术栈得以在现代基础设施中继续演进，并提供预编译二进制文件与完整源代码。
Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability
AI45Researchhttps://github.com/Nebularaid2000/rethink_sft_generalization
重新思考推理监督微调中的泛化：关于优化、数据和模型能力的条件分析
A prevailing narrative in LLM post-training holds that supervised finetuning (SFT) memorizes while reinforcement learning (RL) generalizes. We revisit this claim for reasoning SFT with long chain-of-thought (CoT) supervision and find that cross-domain generalization is not absent but conditional, jointly shaped by optimization dynamics, training data, and base-model capability. Some reported failures are under-optimization artifacts: cross-domain performance first degrades before recovering and improving with extended training (a dip-and-recovery pattern), so shorttraining checkpoints can underestimate generalization. Data quality and structure both matter: low-quality solutions broadly hurt generalization,while verified long-CoT traces yield consistent cross-domain gains. Model capability is essential: stronger models internalize transferable procedural patterns (e.g., backtracking) even from a toy arithmetic game, while weaker ones imitate surface verbosity. This generalization is asymmetric, however: reasoning improves while safety degrades, reframing the question from whether reasoning SFT generalizes to under what conditions and at what cost.
大语言模型后训练领域流行一种观点：监督微调负责记忆，而强化学习负责泛化。我们在长思维链监督的推理监督微调中重新审视这一说法，发现跨领域泛化并非不存在，而是具有条件性，并由优化动态、训练数据和基础模型能力共同塑造。一些已有研究报告的失败其实是优化不足造成的假象：跨领域性能会先下降，随后随着训练延长而恢复并提升，呈现“下降—恢复”模式，因此训练较短的检查点可能低估泛化能力。数据质量和结构同样重要：低质量解答会普遍损害泛化，而经过验证的长思维链轨迹则能稳定带来跨领域收益。模型能力也是关键因素：强模型即使只从一个简单算术游戏中训练，也能内化可迁移的回溯等程序性模式；弱模型则只会模仿表面的冗长表达。不过，这种泛化并不对称：推理能力提升的同时，安全性会下降。因此，问题不再是推理监督微调是否能够泛化，而是它在什么条件下、以何种代价实现泛化。
SkillClaw: Let Skills Evolve Collectively with Agentic Evolver
SkillClaw：让技能借助智能体演化器实现集体演化
Large language model (LLM) agents such as OpenClaw rely on reusable skills to perform complex tasks, yet these skills remain largely static after deployment. As a result, similar workflows, tool usage patterns, and failure modes are repeatedly rediscovered across users, preventing the system from improving with experience. While interactions from different users provide complementary signals about when a skill works or fails, existing systems lack a mechanism to convert such heterogeneous experiences into reliable skill updates. To address these issues, we present SkillClaw, a framework for collective skill evolution in multi-user agent ecosystems, which treats cross-user and over-time interactions as the primary signal for improving skills. SkillClaw continuously aggregates trajectories generated during use and processes them with an autonomous evolver, which identifies recurring behavioral patterns and translates them into updates to the skill set by refining existing skills or extending them with new capabilities. The resulting skills are maintained in a shared repository and synchronized across users, allowing improvements discovered in one context to propagate system-wide while requiring no additional effort from users. By integrating multi-user experience into ongoing skill updates, SkillClaw enables cross-user knowledge transfer and cumulative capability improvement, and experiments on WildClawBench show that limited interaction and feedback, it significantly improves the performance of Qwen3-Max in real-world agent scenarios.
OpenClaw 等大语言模型智能体依靠可复用技能执行复杂任务，但这些技能在部署后基本保持静态。因此，相似的工作流、工具使用模式和失败方式会在不同用户之间被反复重新发现，使系统无法从经验中持续改进。不同用户的交互虽然能提供技能何时有效或失效的互补信号，现有系统却缺少将这些异构经验转化为可靠技能更新的机制。为此，我们提出 SkillClaw，一个面向多用户智能体生态的集体技能演化框架，将跨用户、跨时间的交互作为改进技能的主要信号。SkillClaw 持续汇总使用过程中产生的轨迹，并由自主演化器处理这些轨迹；演化器识别反复出现的行为模式，再通过优化现有技能或扩展新能力，将其转化为技能集合的更新。更新后的技能保存在共享仓库中并同步给所有用户，使一个场景中发现的改进无需用户付出额外努力即可传播至整个系统。通过将多用户经验纳入持续技能更新，SkillClaw 实现了跨用户知识迁移和能力累积提升。WildClawBench 实验表明，只需有限的交互和反馈，它就能显著提高 Qwen3-Max 在真实智能体场景中的表现。
Recursive Multi-Agent Systems
斯坦福大学
递归多智能体系统
Recursive or looped language models have recently emerged as a new scaling axis by iteratively refining the same model computation over latent states to deepen reasoning. We extend such scaling principle from a single model to multi-agent systems, and ask: Can agent collaboration itself be scaled through recursion? To this end, we introduce RecursiveMAS, a recursive multi-agent framework that casts the entire system as a unified latent-space recursive computation. RecursiveMAS connects heterogeneous agents as a collaboration loop through the lightweight RecursiveLink module, enabling in-distribution latent thoughts generation and cross-agent latent state transfer. To optimize our framework, we develop an inner-outer loop learning algorithm for iterative whole-system co-optimization through shared gradient-based credit assignment across recursion rounds. Theoretical analyses of runtime complexity and learning dynamics establish that RecursiveMAS is more efficient than standard text-based MAS and maintains stable gradients during recursive training. Empirically, we instantiate RecursiveMAS under 4 representative agent collaboration patterns and evaluate across 9 benchmarks spanning mathematics, science, medicine, search, and code generation. In comparison with advanced single/multi-agent and recursive computation baselines, RecursiveMAS consistently delivers an average accuracy improvement of 8.3%, together with 1.2times-2.4times end-to-end inference speedup, and 34.6%-75.6% token usage reduction. Code and Data are provided in https://recursivemas.github.io.
递归或循环语言模型近来成为一种新的扩展维度：它在潜在状态上迭代优化同一个模型计算，以加深推理。我们将这一扩展原则从单模型推广到多智能体系统，并提出问题：智能体协作本身能否通过递归进行扩展？为此，我们提出 RecursiveMAS，一个将整个系统表示为统一潜在空间递归计算的递归多智能体框架。RecursiveMAS 通过轻量级 RecursiveLink 模块将异构智能体连接成协作循环，从而生成分布内潜在思维并在智能体之间传递潜在状态。为优化该框架，我们开发了一种内外循环学习算法，通过跨递归轮次共享的基于梯度的信用分配，对整个系统进行迭代联合优化。对运行复杂度和学习动态的理论分析表明，RecursiveMAS 比标准文本多智能体系统更高效，并能在递归训练期间保持梯度稳定。我们在 4 种代表性智能体协作模式下实例化 RecursiveMAS，并在覆盖数学、科学、医学、搜索和代码生成的 9 个基准上评测。相比先进的单智能体、多智能体和递归计算基线，RecursiveMAS 平均准确率稳定提升 8.3%，端到端推理加速 1.2 至 2.4 倍，同时减少 34.6% 至 75.6% 的 token 使用量。代码和数据均已公开。
ClawBench: Can AI Agents Complete Everyday Online Tasks?
自然与人工智能实验室
ClawBench：AI 智能体能否完成日常在线任务？
AI agents may be able to automate your inbox, but can they automate other routine aspects of your life? Everyday online tasks offer a realistic yet unsolved testbed for evaluating the next generation of AI agents. To this end, we introduce ClawBench, an evaluation framework of 153 simple tasks that people need to accomplish regularly in their lives and work, spanning 144 live platforms across 15 categories, from completing purchases and booking appointments to submitting job applications. These tasks require demanding capabilities beyond existing benchmarks, such as obtaining relevant information from user-provided documents, navigating multi-step workflows across diverse platforms, and write-heavy operations like filling in many detailed forms correctly. Unlike existing benchmarks that evaluate agents in offline sandboxes with static pages, ClawBench operates on production websites, preserving the full complexity, dynamic nature, and challenges of real-world web interaction. A lightweight interception layer captures and blocks only the final submission request, ensuring safe evaluation without real-world side effects. Our evaluations of 7 frontier models show that both proprietary and open-source models can complete only a small portion of these tasks. For example, Claude Sonnet 4.6 achieves only 33.3%. Progress on ClawBench brings us closer to AI agents that can function as reliable general-purpose assistants.
AI 智能体或许能够自动处理收件箱，但它们能否自动完成生活中的其他日常事务？日常在线任务为评估下一代 AI 智能体提供了一个真实但尚未解决的测试平台。为此，我们提出 ClawBench，一个由 153 项人们在生活和工作中经常需要完成的简单任务组成的评测框架。这些任务覆盖 15 个类别、144 个真实在线平台，从购物、预约到提交求职申请。它们要求智能体具备超出现有基准的复杂能力，例如从用户提供的文档中获取相关信息、跨不同平台完成多步骤工作流，以及准确填写大量详细表单等重度写入操作。现有基准通常在包含静态页面的离线沙盒中评测智能体，而 ClawBench 直接运行于生产网站，保留真实网络交互的完整复杂性、动态性和挑战。轻量级拦截层只捕获并阻止最终提交请求，从而在不产生现实副作用的前提下保证安全评测。对 7 个前沿模型的评测显示，无论闭源还是开源模型，都只能完成其中很小一部分任务；例如 Claude Sonnet 4.6 的完成率仅为 33.3%。推动 ClawBench 上的进展，将使我们更接近能够充当可靠通用助手的 AI 智能体。
Tstars-Tryon 1.0: Robust and Realistic Virtual Try-On for Diverse Fashion Items
阿里https://huggingface.co/datasets/TaobaoTmall-AlgorithmProducts/Tstars-VTON
Tstars-Tryon 1.0：面向多样时尚单品的稳健逼真虚拟试穿
Recent advances in image generation and editing have opened new opportunities for virtual try-on. However, existing methods still struggle to meet complex real-world demands. We present Tstars-Tryon 1.0, a commercial-scale virtual try-on system that is robust, realistic, versatile, and highly efficient. First, our system maintains a high success rate across challenging cases like extreme poses, severe illumination variations, motion blur, and other in-the-wild conditions. Second, it delivers highly photorealistic results with fine-grained details, faithfully preserving garment texture, material properties, and structural characteristics, while largely avoiding common AI-generated artifacts. Third, beyond apparel try-on, our model supports flexible multi-image composition (up to 6 reference images) across 8 fashion categories, with coordinated control over person identity and background. Fourth, to overcome the latency bottlenecks of commercial deployment, our system is heavily optimized for inference speed, delivering near real-time generation for a seamless user experience. These capabilities are enabled by an integrated system design spanning end-to-end model architecture, a scalable data engine, robust infrastructure, and a multi-stage training paradigm. Extensive evaluation and large-scale product deployment demonstrate that Tstars-Tryon1.0 achieves leading overall performance. To support future research, we also release a comprehensive benchmark. The model has been deployed at an industrial scale on the Taobao App, serving millions of users with tens of millions of requests.
图像生成和编辑技术的近期进展为虚拟试穿带来了新机遇，但现有方法仍难以满足复杂的真实需求。我们提出 Tstars-Tryon 1.0，一个兼具稳健性、真实感、通用性和高效率的商业规模虚拟试穿系统。首先，该系统在极端姿态、剧烈光照变化、运动模糊等开放环境挑战案例中仍能保持较高成功率。其次，它能生成具有精细细节的高度照片级真实结果，在忠实保留服装纹理、材质属性和结构特征的同时，大幅减少常见的 AI 生成伪影。第三，除服装试穿外，模型还支持覆盖 8 类时尚单品、最多使用 6 张参考图像的灵活多图合成，并能协同控制人物身份和背景。第四，为解决商业部署中的延迟瓶颈，系统针对推理速度进行了深度优化，可实现接近实时的生成体验。这些能力来自覆盖端到端模型架构、可扩展数据引擎、稳健基础设施和多阶段训练范式的一体化系统设计。全面评测与大规模产品部署表明，Tstars-Tryon 1.0 的综合性能处于领先水平。我们还发布了一个完整基准以支持后续研究。该模型已在淘宝 App 上实现工业规模部署，为数百万用户处理了数千万次请求。
WildDet3D: Scaling Promptable 3D Detection in the Wild
Ai2
WildDet3D：扩展开放环境中的可提示三维检测
Understanding objects in 3D from a single image is a cornerstone of spatial intelligence. A key step toward this goal is monocular 3D object detection--recovering the extent, location, and orientation of objects from an input RGB image. To be practical in the open world, such a detector must generalize beyond closed-set categories, support diverse prompt modalities, and leverage geometric cues when available. Progress is hampered by two bottlenecks: existing methods are designed for a single prompt type and lack a mechanism to incorporate additional geometric cues, and current 3D datasets cover only narrow categories in controlled environments, limiting open-world transfer. In this work we address both gaps. First, we introduce WildDet3D, a unified geometry-aware architecture that natively accepts text, point, and box prompts and can incorporate auxiliary depth signals at inference time. Second, we present WildDet3D-Data, the largest open 3D detection dataset to date, constructed by generating candidate 3D boxes from existing 2D annotations and retaining only human-verified ones, yielding over 1M images across 13.5K categories in diverse real-world scenes. WildDet3D establishes a new state-of-the-art across multiple benchmarks and settings. In the open-world setting, it achieves 22.6/24.8 AP3D on our newly introduced WildDet3D-Bench with text and box prompts. On Omni3D, it reaches 34.2/36.4 AP3D with text and box prompts, respectively. In zero-shot evaluation, it achieves 40.3/48.9 ODS on Argoverse 2 and ScanNet. Notably, incorporating depth cues at inference time yields substantial additional gains (+20.7 AP on average across settings).
从单张图像理解三维物体是空间智能的基石。实现这一目标的关键步骤是单目三维目标检测，即从输入 RGB 图像中恢复物体的范围、位置和朝向。要在开放世界中具有实用价值，这类检测器必须能够泛化到封闭类别集合之外、支持多种提示模态，并在可用时利用几何线索。当前进展受到两个瓶颈限制：现有方法通常只面向单一提示类型，缺少引入额外几何线索的机制；现有三维数据集又只覆盖受控环境中的少量类别，限制了开放世界迁移。本文同时解决这两个问题。首先，我们提出 WildDet3D，一个统一的几何感知架构，原生支持文本、点和框提示，并可在推理时融合辅助深度信号。其次，我们构建 WildDet3D-Data，这是迄今最大的开放三维检测数据集：从现有二维标注生成候选三维框，只保留经人工验证的结果，最终覆盖多样真实场景中的 13500 个类别和超过 100 万张图像。WildDet3D 在多个基准和设置上刷新了最佳性能。在开放世界设置中，使用文本和框提示时，它在新提出的 WildDet3D-Bench 上分别达到 22.6 和 24.8 AP3D；在 Omni3D 上分别达到 34.2 和 36.4 AP3D；零样本评测中，它在 Argoverse 2 和 ScanNet 上分别达到 40.3 和 48.9 ODS。尤其值得注意的是，在推理时加入深度线索可在各项设置上平均额外提升 20.7 AP。
LLaDA2.0-Uni: Unifying Multimodal Understanding and Generation with Diffusion Large Language Model
inclusionAI
LLaDA2.0-Uni：用扩散大语言模型统一多模态理解与生成
We present LLaDA2.0-Uni, a unified discrete diffusion large language model (dLLM) that supports multimodal understanding and generation within a natively integrated framework. Its architecture combines a fully semantic discrete tokenizer, a MoE-based dLLM backbone, and a diffusion decoder. By discretizing continuous visual inputs via SigLIP-VQ, the model enables block-level masked diffusion for both text and vision inputs within the backbone, while the decoder reconstructs visual tokens into high-fidelity images. Inference efficiency is enhanced beyond parallel decoding through prefix-aware optimizations in the backbone and few-step distillation in the decoder. Supported by carefully curated large-scale data and a tailored multi-stage training pipeline, LLaDA2.0-Uni matches specialized VLMs in multimodal understanding while delivering strong performance in image generation and editing. Its native support for interleaved generation and reasoning establishes a promising and scalable paradigm for next-generation unified foundation models. Codes and models are available at https://github.com/inclusionAI/LLaDA2.0-Uni.
我们提出 LLaDA2.0-Uni，一个在原生一体化框架中同时支持多模态理解与生成的统一离散扩散大语言模型。其架构结合全语义离散 tokenizer、基于混合专家的扩散大语言模型骨干网络和扩散解码器。模型通过 SigLIP-VQ 将连续视觉输入离散化，使骨干网络能够同时对文本和视觉输入执行块级掩码扩散，随后由解码器将视觉 token 重建为高保真图像。除并行解码外，骨干网络中的前缀感知优化和解码器中的少步蒸馏进一步提高了推理效率。在精心筛选的大规模数据和定制多阶段训练流程支持下，LLaDA2.0-Uni 的多模态理解能力可媲美专用视觉语言模型，同时在图像生成与编辑方面表现强劲。它原生支持交织式生成与推理，为下一代统一基础模型建立了一种前景良好且可扩展的范式。代码和模型均已公开。
InCoder-32B-Thinking: Industrial Code World Model for Thinking
InCoder-32B-Thinking：面向推理的工业代码世界模型
Industrial software development across chip design, GPU optimization, and embedded systems lacks expert reasoning traces showing how engineers reason about hardware constraints and timing semantics. In this work, we propose InCoder-32B-Thinking, trained on the data from the Error-driven Chain-of-Thought (ECoT) synthesis framework with an industrial code world model (ICWM) to generate reasoning traces. Specifically, ECoT generates reasoning chains by synthesizing the thinking content from multi-turn dialogue with environmental error feedback, explicitly modeling the error-correction process. ICWM is trained on domain-specific execution traces from Verilog simulation, GPU profiling, etc., learns the causal dynamics of how code affects hardware behavior, and enables self-verification by predicting execution outcomes before actual compilation. All synthesized reasoning traces are validated through domain toolchains, creating training data matching the natural reasoning depth distribution of industrial tasks. Evaluation on 14 general (81.3% on LiveCodeBench v5) and 9 industrial benchmarks (84.0% in CAD-Coder and 38.0% on KernelBench) shows InCoder-32B-Thinking achieves top-tier open-source results across all domains.GPU Optimization
芯片设计、GPU 优化和嵌入式系统等工业软件开发领域缺少专家推理轨迹，无法展现工程师如何围绕硬件约束和时序语义展开推理。本文提出 InCoder-32B-Thinking，它使用错误驱动思维链（ECoT）合成框架生成的数据进行训练，并借助工业代码世界模型（ICWM）生成推理轨迹。具体而言，ECoT 从带有环境错误反馈的多轮对话中合成思考内容以生成推理链，从而显式建模错误修正过程。ICWM 在 Verilog 仿真、GPU 性能分析等领域专用执行轨迹上训练，学习代码如何影响硬件行为的因果动态，并通过在实际编译前预测执行结果实现自我验证。所有合成推理轨迹都经过领域工具链验证，从而形成符合工业任务自然推理深度分布的训练数据。在 14 个通用基准和 9 个工业基准上的评测显示，InCoder-32B-Thinking 在所有领域均取得顶尖开源成绩，包括 LiveCodeBench v5 上的 81.3%、CAD-Coder 上的 84.0% 和 KernelBench 上的 38.0%。
Video-MME-v2: Towards the Next Stage in Benchmarks for Comprehensive Video Understanding
MME-Benchmarks
Video-MME-v2：迈向综合视频理解基准的下一阶段
With the rapid advancement of video understanding, existing benchmarks are becoming increasingly saturated, exposing a critical discrepancy between inflated leaderboard scores and real-world model capabilities. To address this widening gap, we introduce Video-MME-v2, a comprehensive benchmark designed to rigorously evaluate the robustness and faithfulness of video understanding. To systematically evaluate model capabilities, we design a progressive tri-level hierarchy that incrementally increases the complexity of video comprehension, ranging from multi-point visual information aggregation, to temporal dynamics modeling, and ultimately to complex multimodal reasoning. Besides, in contrast to conventional per-question accuracy, we propose a group-based non-linear evaluation strategy that enforces both consistency across related queries and coherence in multi-step reasoning. It penalizes fragmented or guess-based correctness and assigns credit only to answers supported by valid reasoning. To guarantee data quality, Video-MME-v2 is constructed through a rigorously controlled human annotation pipeline, involving 12 annotators and 50 independent reviewers. Backed by 3,300 human-hours and up to 5 rounds of quality assurance, Video-MME-v2 aims to serve as one of the most authoritative video benchmarks. Extensive experiments reveal a substantial gap between current best model Gemini-3-Pro and human experts, and uncover a clear hierarchical bottleneck where errors in visual information aggregation and temporal modeling propagate to limit high-level reasoning. We further find that thinking-based reasoning is highly dependent on textual cues, improving performance with subtitles but sometimes degrading it in purely visual settings. By exposing these limitations, Video-MME-v2 establishes a demanding new testbed for the development of next-generation video MLLMs.
随着视频理解快速发展，现有基准日趋饱和，暴露出排行榜高分与模型真实能力之间的严重偏差。为缩小这一不断扩大的差距，我们提出 Video-MME-v2，一个旨在严格评估视频理解稳健性和忠实度的综合基准。为系统评测模型能力，我们设计了一个难度逐步提升的三级层次结构：从多点视觉信息聚合，到时间动态建模，最终到复杂多模态推理。不同于传统的逐题准确率，我们还提出基于问题组的非线性评测策略，同时要求相关问题之间保持一致，并保证多步推理连贯。该策略会惩罚碎片化或依赖猜测的正确答案，只为有有效推理支持的答案计分。为保证数据质量，Video-MME-v2 采用严格控制的人工标注流程构建，共有 12 名标注员和 50 名独立审核员参与。项目投入 3300 个人工小时并进行最多 5 轮质量保证，目标是成为最权威的视频基准之一。大量实验揭示了当前最佳模型 Gemini-3-Pro 与人类专家之间的显著差距，也发现了清晰的层级瓶颈：视觉信息聚合和时间建模中的错误会向上传播，限制高级推理。我们还发现，思考式推理高度依赖文本线索；字幕可提升表现，但在纯视觉设置中有时反而会降低性能。通过揭示这些限制，Video-MME-v2 为下一代视频多模态大模型提供了一个高难度新测试平台。
Agentic World Modeling: Foundations, Capabilities, Laws, and Beyond
https://github.com/matrix-agent/awesome-agentic-world-modeling
智能体世界建模：基础、能力、规律及未来
As AI systems move from generating text to accomplishing goals through sustained interaction, the ability to model environment dynamics becomes a central bottleneck. Agents that manipulate objects, navigate software, coordinate with others, or design experiments require predictive environment models, yet the term world model carries different meanings across research communities. We introduce a "levels x laws" taxonomy organized along two axes. The first defines three capability levels: L1 Predictor, which learns one-step local transition operators; L2 Simulator, which composes them into multi-step, action-conditioned rollouts that respect domain laws; and L3 Evolver, which autonomously revises its own model when predictions fail against new evidence. The second identifies four governing-law regimes: physical, digital, social, and scientific. These regimes determine what constraints a world model must satisfy and where it is most likely to fail. Using this framework, we synthesize over 400 works and summarize more than 100 representative systems spanning model-based reinforcement learning, video generation, web and GUI agents, multi-agent social simulation, and AI-driven scientific discovery. We analyze methods, failure modes, and evaluation practices across level-regime pairs, propose decision-centric evaluation principles and a minimal reproducible evaluation package, and outline architectural guidance, open problems, and governance challenges. The resulting roadmap connects previously isolated communities and charts a path from passive next-step prediction toward world models that can simulate, and ultimately reshape, the environments in which agents operate.
随着 AI 系统从生成文本转向通过持续交互完成目标，建模环境动态的能力成为核心瓶颈。操纵物体、操作软件、与他者协作或设计实验的智能体都需要预测式环境模型，但“世界模型”在不同研究社区中含义各异。我们提出一个沿两条轴线组织的“层级 × 规律”分类体系。第一条轴线定义三个能力层级：L1 预测器学习单步局部转移算子；L2 仿真器将这些算子组合为遵守领域规律、以动作为条件的多步轨迹；L3 演化器则在预测与新证据冲突时自主修正自身模型。第二条轴线识别四种支配规律：物理、数字、社会和科学规律。这些规律决定世界模型必须满足哪些约束，以及它最可能在哪里失败。借助该框架，我们综合分析了 400 余项工作，并总结超过 100 个代表性系统，覆盖基于模型的强化学习、视频生成、网页与 GUI 智能体、多智能体社会仿真以及 AI 驱动的科学发现。我们分析不同层级与规律组合下的方法、失败模式和评测实践，提出以决策为中心的评测原则与最小可复现评测包，并概述架构指导、开放问题和治理挑战。该路线图连接了此前彼此分离的研究社区，并勾勒出一条从被动预测下一步，走向能够模拟乃至重塑智能体所处环境的世界模型发展路径。
OpenWorldLib: A Unified Codebase and Definition of Advanced World Models
https://wcny4qa9krto.feishu.cn/wiki/XtPJwf5XQipP7RkeVv0ckyWlnNd
北京大学
OpenWorldLib：高级世界模型的统一代码库与定义
World models have garnered significant attention as a promising research direction in artificial intelligence, yet a clear and unified definition remains lacking. In this paper, we introduce OpenWorldLib, a comprehensive and standardized inference framework for Advanced World Models. Drawing on the evolution of world models, we propose a clear definition: a world model is a model or framework centered on perception, equipped with interaction and long-term memory capabilities, for understanding and predicting the complex world. We further systematically categorize the essential capabilities of world models. Based on this definition, OpenWorldLib integrates models across different tasks within a unified framework, enabling efficient reuse and collaborative inference. Finally, we present additional reflections and analyses on potential future directions for world model research. Code link: https://github.com/OpenDCAI/OpenWorldLib
世界模型作为人工智能中一个前景广阔的研究方向受到广泛关注，但目前仍缺乏清晰统一的定义。本文提出 OpenWorldLib，一个面向高级世界模型的综合标准化推理框架。基于世界模型的发展历程，我们给出明确定义：世界模型是以感知为中心、具备交互和长期记忆能力，用于理解并预测复杂世界的模型或框架。我们进一步系统梳理了世界模型应具备的核心能力。在这一定义基础上，OpenWorldLib 将面向不同任务的模型集成进统一框架，实现高效复用与协同推理。最后，我们还对世界模型研究潜在的未来方向进行了进一步思考与分析。代码已经公开。