每日論文

我們推出 Ling 2.0——一個基於「每次激活皆提升推理能力」原則構建的系列化推理導向語言基礎模型。該系列在統一混合專家模型框架下，設計參數規模可從數百億擴展至一萬億，並憑藉實證縮放定律實現高稀疏度、跨尺度一致性與高效能特性。系列包含三款非思維（指令）模型：Ling-mini-2.0、Ling-flash-2.0 與 Ling-1T，總參數規模從160億至1萬億，相比稠密模型實現最高7倍激活計算效率。Ling 2.0 整合了模型架構、預訓練、後訓練及基礎設施的協同創新：採用具備多令牌預測的高稀疏MoE架構以實現高效推理，融合推理導向數據與訓練中期思維鏈激活，實施基於強化的精調技術（DFT、Evo-CoT），並實現全尺度FP8訓練與細粒度異構流水線。在萬億參數級別，Ling-1T 建立了推理準確率與計算效率的新帕累托前沿，證明稀疏激活機制與推理目標精準對齊時，可實現可擴展的高效智能。整體而言，Ling 2.0 為推進未來推理與思維模型（包括基於同底座的Ring系列）提供了連貫、開放且高效的基礎框架。

以生成模型（如擴散策略）為參數的隱式策略，已成為機器人領域中策略學習與視覺-語言-動作模型的主流方法。然而，這類方法常面臨計算成本高、曝光偏差及推理動態不穩定等問題，導致在分佈偏移下出現發散現象。基於能量的模型通過端到端學習能量景觀並建模平衡動力學，能有效改善魯棒性並降低曝光偏差，但傳統上以EBM參數化的策略難以擴展至複雜場景。近期提出的能量型Transformer證明了EBM在高維空間的可擴展性，但其在物理實體模型核心挑戰中的應用潛力尚未充分探索。我們提出新型能量架構EBT-Policy，專注解決機器人與真實世界環境的關鍵問題。在仿真與實物任務中，EBT-Policy不僅持續超越基於擴散的策略，更顯著降低訓練與推理計算量——部分任務僅需兩次推理步驟即可收斂，較擴散策略的100步實現50倍壓縮。尤為突出的是，EBT-Policy展現出前所未有的湧現能力：僅通過行為克隆且無需重試訓練，即可實現失敗動作序列的零樣本恢復。通過利用標量能量進行不確定性感知推理與動態計算資源分配，EBT-Policy為分佈偏移下實現魯棒、可泛化的機器人行為開闢了新路徑。

測試時擴展（TTS）技術透過在推理階段分配額外計算資源來提升大型語言模型（LLM）的性能，通常採用並行、序列或混合擴展方式。然而，現有研究往往預設固定的協作架構（如拓撲結構）和單一模型使用模式，忽略了最優架構與模型組合會隨任務不同而變化的特性。為此，我們研究在固定計算預算下，為TTS搜尋計算最優的模型組合與架構這一新問題。我們將其形式化為多LLM協作圖結構：節點編碼角色與LLM模型分配，邊緣捕捉資訊流動。該問題的挑戰在於：（i）組合搜尋空間過於龐大；（ii）任務特定需求需要定制化設計。為解決這些難題，我們將問題重新表述為概率圖優化，並透過預實驗歸納出TTS協作圖的三項實證規律。基於這些規律，我們提出Agent-REINFORCE框架——該框架透過LLM智能體增強，將REINFORCE演算法的「採樣-梯度-更新」流程映射為「採樣-反饋-更新」，其中文字化反饋作為梯度替代品來更新概率圖，從而高效搜尋最優多LLM協作圖。實驗表明，Agent-REINFORCE在樣本效率和搜尋性能上均優於傳統基線與LLM基線，並能有效在準確率與推理延遲的聯合目標下識別最優圖結構。

我們推出[Cosmos-Predict2.5]——宇宙世界物理人工智慧基礎模型的最新世代。基於流式架構構建的[Cosmos-Predict2.5]將文字生成世界、圖像生成世界與影片生成世界三大功能整合於單一模型中，並運用物理AI視覺語言模型[Cosmos-Reason1]提供更豐富的文本錨定與更精細的世界模擬控制。該模型在2億支精選影片片段上進行訓練，並透過強化學習式後訓練優化，相較[Cosmos-Predict1]在影片品質與指令遵循方面實現顯著提升，現提供20億與140億參數規模的版本。這些突破性能力可為機器人與自主系統提供更可靠的合成數據生成、策略評估及閉環模擬。我們同時推出控制網路風格的[Cosmos-Transfer2.5]框架，用於模擬到現實與現實到現實的世界轉譯。儘管體積僅為[Cosmos-Transfer1]的1/3.5，其仍能實現更高保真度與強健的長時序影片生成。這些進展共同確立了[Cosmos-Predict2.5]與[Cosmos-Transfer2.5]作為擴展具身智能的通用工具地位。為加速物理AI領域的研究與部署，我們於https://github.com/nvidia-cosmos/cosmos-predict2.5 及 https://github.com/nvidia-cosmos/cosmos-transfer2.5 開源原始碼、預訓練模型與精選基準測試集（採用NVIDIA開放模型許可協議）。期待這些開放資源能降低技術門檻，推動下一代具身智能建設的創新發展。

近期多模態生成模型的進展顯著提升了圖像編輯能力。然而，現有生成模型在處理需要隱式推理的多樣化複雜圖像編輯任務時仍面臨挑戰，這凸顯了建立系統性評估模型在不同推理場景下表現的綜合基準的必要性。現有基準主要聚焦於現實場景中的單一物件屬性轉換，雖具實效性卻存在兩大關鍵侷限：（1）普遍忽略多物件互動及涉及人為規則的遊戲世界場景，而這類場景在實際應用中十分常見；（2）僅依賴文本參照評估生成圖像，可能在複雜推理場景中導致系統性誤判。為此，本研究提出統一推理式圖像編輯評估基準UniREditBench，包含2,700個精心策劃的樣本，涵蓋現實與遊戲世界場景的8個主維度及18個子維度。為提升評估可靠性，我們引入多模態雙參照評估機制，為每個樣本提供文本與真實圖像雙重參照。此外，我們設計了自動化多場景數據合成流程，構建包含高質量思維鏈推理標註的大規模合成數據集UniREdit-Data-100K。通過在該數據集上微調Bagel模型，我們開發出UniREdit-Bagel，其在領域內與分佈外場景均展現顯著性能提升。透過對開源與閉源圖像編輯模型的全面基準測試，我們揭示了各模型在不同維度的優劣勢。

圖神經網絡透過自下而上的訊息傳遞機制運作，這種方式與人類視覺感知存在根本差異——後者會直覺性地先捕捉全局結構。我們探索了視覺模型在圖理解領域尚未被充分發掘的潛力，發現其在經典基準測試中能達到與圖神經網絡相當的性能，同時展現出截然不同的學習模式。這種行為差異，加上現有基準測試中存在的領域特徵與拓撲理解相互混淆的局限性，促使我們推出GraphAbstract基準。該基準評估模型像人類一樣感知全局圖屬性的能力：識別組織原型、檢測對稱性、感知連接強度以及定位關鍵元素。實驗結果表明，在需要整體結構理解的任務上，視覺模型顯著優於圖神經網絡，並能保持跨圖規模的泛化能力；而圖神經網絡則在全局模式抽象方面表現掙扎，且性能隨圖規模增大而衰減。本研究證實視覺模型具有卓越但未被充分利用的圖結構理解能力，特別是在需要全局拓撲感知和尺度不變推理的問題上。這些發現為開發更有效的圖基礎模型開闢了新途徑，尤其適用於以整體模式識別為主導的任務場景。

重光照技術是一項兼具實用需求與藝術價值的關鍵任務，近期擴散模型通過實現豐富可控的照明效果展現出強大潛力。然而，由於這類模型通常在語義潛空間進行優化，而潛空間中的鄰近性無法保證視覺空間的物理正確性，因此常產生不真實的結果，例如過曝的高光、錯位的陰影與錯誤的遮擋。為此我們提出UniLumos——一個適用於圖像與視頻的統一重光照框架，將RGB空間的幾何反饋引入流匹配主幹網絡。通過從模型輸出中提取深度圖和法線圖進行監督，我們顯式地將照明效果與場景結構對齊，從而提升物理合理性。但這種反饋機制需要高質量輸出作為視覺空間的監督信號，導致傳統多步去噪方法計算成本高昂。為緩解此問題，我們採用路徑一致性學習，使監督在少步數訓練機制下仍保持有效性。為實現細粒度重光照控制與監督，我們設計了結構化的六維標註協議，用於捕捉核心光照屬性。基於此協議，我們提出LumosBench解耦屬性級評測基準，通過大型視覺語言模型評估光照可控性，實現跨維度的自動化可解釋重光照精度評估。大量實驗表明，UniLumos在實現顯著提升物理一致性的同時，達到了業界頂尖的重光照質量，並為圖像和視頻重光照帶來20倍加速效果。代碼已開源於：https://github.com/alibaba-damo-academy/Lumos-Custom。

Large reasoning models (LRMs) show strong capabilities in complex reasoning, yet their marginal gains on evidence-dependent factual questions are limited. We find this limitation is partially attributable to a reasoning-answer hit gap, where the model identifies the correct facts during reasoning but fails to incorporate them into the final response, thereby reducing factual fidelity. To address this issue, we propose MR-ALIGN, a Meta-Reasoning informed alignment framework that enhances factuality without relying on external verifiers. MR-ALIGN quantifies state transition probabilities along the model's thinking process and constructs a transition-aware implicit reward that reinforces beneficial reasoning patterns while suppressing defective ones at the atomic thinking segments. This re-weighting reshapes token-level signals into probability-aware segment scores, encouraging coherent reasoning trajectories that are more conducive to factual correctness. Empirical evaluations across four factual QA datasets and one long-form factuality benchmark show that MR-ALIGN consistently improves accuracy and truthfulness while reducing misleading reasoning. These results highlight that aligning the reasoning process itself, rather than merely the outputs, is pivotal for advancing factuality in LRMs.

Unified multimodal models (UMMs) have emerged as a powerful paradigm for seamlessly unifying text and image understanding and generation. However, prevailing evaluations treat these abilities in isolation, such that tasks with multimodal inputs and outputs are scored primarily through unimodal reasoning, i.e., textual benchmarks emphasize language-based reasoning, while visual benchmarks emphasize reasoning outcomes manifested in the pixels. We introduce ROVER to address this pressing need to test reciprocal cross-modal reasoning, the use of one modality to guide, verify, or refine outputs in the other, an ability central to the vision of unified multimodal intelligence. ROVER is a human-annotated benchmark that explicitly targets reciprocal cross-modal reasoning, which contains 1312 tasks grounded in 1876 images, spanning two complementary settings. Verbally-augmented reasoning for visual generation evaluates whether models can use verbal prompts and reasoning chains to guide faithful image synthesis. Visually-augmented reasoning for verbal generation evaluates whether models can generate intermediate visualizations that strengthen their own reasoning processes for question answering. Experiments on 17 unified models reveal two key findings: (i) Cross-modal reasoning determines visual generation quality, with interleaved models significantly outperforming non-interleaved ones; notably, combining strong unimodal models fails to achieve comparable reasoning. (ii) Models show dissociation between physical and symbolic reasoning: they succeed at interpreting perceptual concepts literally but fail to construct visual abstractions for symbolic tasks, where faulty reasoning harms performance. These results highlight reciprocal cross-modal reasoning as a critical frontier for enabling true omnimodal generation.

Motion imitation is a promising approach for humanoid locomotion, enabling agents to acquire humanlike behaviors. Existing methods typically rely on high-quality motion capture datasets such as AMASS, but these are scarce and expensive, limiting scalability and diversity. Recent studies attempt to scale data collection by converting large-scale internet videos, exemplified by Humanoid-X. However, they often introduce physical artifacts such as floating, penetration, and foot skating, which hinder stable imitation. In response, we introduce PHUMA, a Physically-grounded HUMAnoid locomotion dataset that leverages human video at scale, while addressing physical artifacts through careful data curation and physics-constrained retargeting. PHUMA enforces joint limits, ensures ground contact, and eliminates foot skating, producing motions that are both large-scale and physically reliable. We evaluated PHUMA in two sets of conditions: (i) imitation of unseen motion from self-recorded test videos and (ii) path following with pelvis-only guidance. In both cases, PHUMA-trained policies outperform Humanoid-X and AMASS, achieving significant gains in imitating diverse motions. The code is available at https://davian-robotics.github.io/PHUMA.

Current motion-conditioned video generation methods suffer from prohibitive latency (minutes per video) and non-causal processing that prevents real-time interaction. We present MotionStream, enabling sub-second latency with up to 29 FPS streaming generation on a single GPU. Our approach begins by augmenting a text-to-video model with motion control, which generates high-quality videos that adhere to the global text prompt and local motion guidance, but does not perform inference on the fly. As such, we distill this bidirectional teacher into a causal student through Self Forcing with Distribution Matching Distillation, enabling real-time streaming inference. Several key challenges arise when generating videos of long, potentially infinite time-horizons: (1) bridging the domain gap from training on finite length and extrapolating to infinite horizons, (2) sustaining high quality by preventing error accumulation, and (3) maintaining fast inference, without incurring growth in computational cost due to increasing context windows. A key to our approach is introducing carefully designed sliding-window causal attention, combined with attention sinks. By incorporating self-rollout with attention sinks and KV cache rolling during training, we properly simulate inference-time extrapolations with a fixed context window, enabling constant-speed generation of arbitrarily long videos. Our models achieve state-of-the-art results in motion following and video quality while being two orders of magnitude faster, uniquely enabling infinite-length streaming. With MotionStream, users can paint trajectories, control cameras, or transfer motion, and see results unfold in real-time, delivering a truly interactive experience.

Recently, large language models (LLMs) have demonstrated remarkable problem-solving capabilities by autonomously integrating with external tools for collaborative reasoning. However, due to the inherently complex and diverse nature of multimodal information, enabling multimodal large language models (MLLMs) to flexibly and efficiently utilize external tools during reasoning remains an underexplored challenge. In this work, we introduce ToolScope, an agentic framework designed to unify global planning with local multimodal perception, adopting a specialized Perceive tool to mitigates visual context degradation in long-horizon VQA task. ToolScope comprises three primary components: the Global Navigator, the Agentic Executor, and the Response Synthesizer. The Global Navigator functions as a "telescope", offering high-level strategic guidance. The Agentic Executor operates iteratively to augment MLLM with local perception through the integration of external tools-Search, Code, and Perceive. Finally, the Response Synthesizer consolidates and organizes the reasoning process into a coherent, user-friendly output. We evaluate ToolScope on four VQA benchmarks across diverse domains, including VQA 2.0, ScienceQA, MAT-Search and MathVista. It demonstrates strong generalization capabilities, achieving an average performance improvement of up to +6.69% across all datasets.

We introduce LongCat-Flash-Omni, a state-of-the-art open-source omni-modal model with 560 billion parameters, excelling at real-time audio-visual interaction. By adopting a curriculum-inspired progressive training strategy that transitions from simpler to increasingly complex modality sequence modeling tasks, LongCat-Flash-Omni attains comprehensive multimodal capabilities while maintaining strong unimodal capability. Building upon LongCat-Flash, which adopts a high-performance Shortcut-connected Mixture-of-Experts (MoE) architecture with zero-computation experts, LongCat-Flash-Omni integrates efficient multimodal perception and speech reconstruction modules. Despite its immense size of 560B parameters (with 27B activated), LongCat-Flash-Omni achieves low-latency real-time audio-visual interaction. For training infrastructure, we developed a modality-decoupled parallelism scheme specifically designed to manage the data and model heterogeneity inherent in large-scale multimodal training. This innovative approach demonstrates exceptional efficiency by sustaining over 90% of the throughput achieved by text-only training. Extensive evaluations show that LongCat-Flash-Omni achieves state-of-the-art performance on omni-modal benchmarks among open-source models. Furthermore, it delivers highly competitive results across a wide range of modality-specific tasks, including text, image, and video understanding, as well as audio understanding and generation. We provide a comprehensive overview of the model architecture design, training procedures, and data strategies, and open-source the model to foster future research and development in the community.

Recent advances in large language model (LLM) reasoning through reinforcement learning rely on annotated datasets for verifiable rewards, which may limit models' ability to surpass human-level performance. While self-play offers a promising alternative, existing approaches depend on external verifiers or cannot learn open-endedly. We present Open-Ended Self-Improving Reasoner (OpenSIR), a self-play framework where an LLM learns to generate and solve novel problems by alternating teacher and student roles without external supervision. To generate novel problems, OpenSIR optimises for both difficulty and diversity, rewarding problems that challenge appropriately while exploring distinct concepts, enabling open-ended mathematical discovery. Starting from a single trivial seed problem, OpenSIR substantially improves instruction models: Llama-3.2-3B-Instruct advances from 73.9 to 78.3 on GSM8K, and from 28.8 to 34.4 on College Math, while Gemma-2-2B-Instruct rises from 38.5 to 58.7 on GSM8K. Our analyses reveal that OpenSIR achieves open-ended learning through co-evolving teacher-student roles that adaptively calibrate difficulty and drive diverse exploration, progressing autonomously from basic to advanced mathematics.

The prevailing video retrieval paradigm is structurally misaligned, as narrow benchmarks incentivize correspondingly limited data and single-task training. Therefore, universal capability is suppressed due to the absence of a diagnostic evaluation that defines and demands multi-dimensional generalization. To break this cycle, we introduce a framework built on the co-design of evaluation, data, and modeling. First, we establish the Universal Video Retrieval Benchmark (UVRB), a suite of 16 datasets designed not only to measure performance but also to diagnose critical capability gaps across tasks and domains. Second, guided by UVRB's diagnostics, we introduce a scalable synthesis workflow that generates 1.55 million high-quality pairs to populate the semantic space required for universality. Finally, we devise the Modality Pyramid, a curriculum that trains our General Video Embedder (GVE) by explicitly leveraging the latent interconnections within our diverse data. Extensive experiments show GVE achieves state-of-the-art zero-shot generalization on UVRB. In particular, our analysis reveals that popular benchmarks are poor predictors of general ability and that partially relevant retrieval is a dominant but overlooked scenario. Overall, our co-designed framework provides a practical path to escape the limited scope and advance toward truly universal video retrieval.

The frontier of visual reasoning is shifting toward models like OpenAI o3, which can intelligently create and operate tools to transform images for problem-solving, also known as thinking-with-images in chain-of-thought. Yet existing benchmarks fail to fully capture this advanced capability. Even Visual Search, the most common benchmark for current thinking-with-images methods, tests only basic operations such as localization and cropping, offering little insight into more complex, dynamic, and tool-dependent reasoning. We introduce TIR-Bench, a comprehensive benchmark for evaluating agentic thinking-with-images across 13 diverse tasks, each requiring novel tool use for image processing and manipulation in chain-of-thought. We evaluate 22 multimodal large language models (MLLMs), from leading open-sourced and proprietary models to those with explicit tool-use augmentation. Results show that TIR-Bench is universally challenging, and strong performance requires genuine thinking-with-images capabilities. Finally, we present a pilot study comparing direct versus agentic fine-tuning.

Vision-language models demonstrate unprecedented performance and generalization across a wide range of tasks and scenarios. Integrating these foundation models into robotic navigation systems opens pathways toward building general-purpose robots. Yet, evaluating these models' navigation capabilities remains constrained by costly real-world trials, overly simplified simulations, and limited benchmarks. We introduce NaviTrace, a high-quality Visual Question Answering benchmark where a model receives an instruction and embodiment type (human, legged robot, wheeled robot, bicycle) and must output a 2D navigation trace in image space. Across 1000 scenarios and more than 3000 expert traces, we systematically evaluate eight state-of-the-art VLMs using a newly introduced semantic-aware trace score. This metric combines Dynamic Time Warping distance, goal endpoint error, and embodiment-conditioned penalties derived from per-pixel semantics and correlates with human preferences. Our evaluation reveals consistent gap to human performance caused by poor spatial grounding and goal localization. NaviTrace establishes a scalable and reproducible benchmark for real-world robotic navigation. The benchmark and leaderboard can be found at https://leggedrobotics.github.io/navitrace_webpage/.

Understanding Rebus Puzzles (Rebus Puzzles use pictures, symbols, and letters to represent words or phrases creatively) requires a variety of skills such as image recognition, cognitive skills, commonsense reasoning, multi-step reasoning, image-based wordplay, etc., making this a challenging task for even current Vision-Language Models. In this paper, we present left|,circlearrowright,text{BUS},right|, a large and diverse benchmark of 1,333 English Rebus Puzzles containing different artistic styles and levels of difficulty, spread across 18 categories such as food, idioms, sports, finance, entertainment, etc. We also propose RebusDescProgICE, a model-agnostic framework which uses a combination of an unstructured description and code-based, structured reasoning, along with better, reasoning-based in-context example selection, improving the performance of Vision-Language Models on left|,circlearrowright,text{BUS},right| by 2.1-4.1% and 20-30% using closed-source and open-source models respectively compared to Chain-of-Thought Reasoning.

Reading measurement instruments is effortless for humans and requires relatively little domain expertise, yet it remains surprisingly challenging for current vision-language models (VLMs) as we find in preliminary evaluation. In this work, we introduce MeasureBench, a benchmark on visual measurement reading covering both real-world and synthesized images of various types of measurements, along with an extensible pipeline for data synthesis. Our pipeline procedurally generates a specified type of gauge with controllable visual appearance, enabling scalable variation in key details such as pointers, scales, fonts, lighting, and clutter. Evaluation on popular proprietary and open-weight VLMs shows that even the strongest frontier VLMs struggle measurement reading in general. A consistent failure mode is indicator localization: models can read digits or labels but misidentify the key positions of pointers or alignments, leading to big numeric errors despite plausible textual reasoning. We have also conducted preliminary experiments with reinforcement learning over synthetic data, and find encouraging results on in-domain synthetic subset but less promising for real-world images. Our analysis highlights a fundamental limitation of current VLMs in fine-grained spatial grounding. We hope this resource can help future advances on visually grounded numeracy and precise spatial perception of VLMs, bridging the gap between recognizing numbers and measuring the world.

We introduce Trove, an easy-to-use open-source retrieval toolkit that simplifies research experiments without sacrificing flexibility or speed. For the first time, we introduce efficient data management features that load and process (filter, select, transform, and combine) retrieval datasets on the fly, with just a few lines of code. This gives users the flexibility to easily experiment with different dataset configurations without the need to compute and store multiple copies of large datasets. Trove is highly customizable: in addition to many built-in options, it allows users to freely modify existing components or replace them entirely with user-defined objects. It also provides a low-code and unified pipeline for evaluation and hard negative mining, which supports multi-node execution without any code changes. Trove's data management features reduce memory consumption by a factor of 2.6. Moreover, Trove's easy-to-use inference pipeline incurs no overhead, and inference times decrease linearly with the number of available nodes. Most importantly, we demonstrate how Trove simplifies retrieval experiments and allows for arbitrary customizations, thus facilitating exploratory research.

Data selection is a critical aspect of Reinforcement Learning with Verifiable Rewards (RLVR) for enhancing the reasoning capabilities of large language models (LLMs). Current data selection methods are largely heuristic-based, lacking theoretical guarantees and generalizability. This work proposes a theoretically-grounded approach using influence functions to estimate the contribution of each data point to the learning objective. To overcome the prohibitive computational cost of policy rollouts required for online influence estimation, we introduce an off-policy influence estimation method that efficiently approximates data influence using pre-collected offline trajectories. Furthermore, to manage the high-dimensional gradients of LLMs, we employ sparse random projection to reduce dimensionality and improve storage and computation efficiency. Leveraging these techniques, we develop Curriculum RL with Off-Policy Influence guidance (CROPI), a multi-stage RL framework that iteratively selects the most influential data for the current policy. Experiments on models up to 7B parameters demonstrate that CROPI significantly accelerates training. On a 1.5B model, it achieves a 2.66x step-level acceleration while using only 10\% of the data per stage compared to full-dataset training. Our results highlight the substantial potential of influence-based data selection for efficient RLVR.

Recent advances in Multimodal Large Language Models (MLLMs) have significantly improved 2D visual understanding, prompting interest in their application to complex 3D reasoning tasks. However, it remains unclear whether these models can effectively capture the detailed spatial information required for robust real-world performance, especially cross-view consistency, a key requirement for accurate 3D reasoning. Considering this issue, we introduce Viewpoint Learning, a task designed to evaluate and improve the spatial reasoning capabilities of MLLMs. We present the Viewpoint-100K dataset, consisting of 100K object-centric image pairs with diverse viewpoints and corresponding question-answer pairs. Our approach employs a two-stage fine-tuning strategy: first, foundational knowledge is injected to the baseline MLLM via Supervised Fine-Tuning (SFT) on Viewpoint-100K, resulting in significant improvements across multiple tasks; second, generalization is enhanced through Reinforcement Learning using the Group Relative Policy Optimization (GRPO) algorithm on a broader set of questions. Additionally, we introduce a hybrid cold-start initialization method designed to simultaneously learn viewpoint representations and maintain coherent reasoning thinking. Experimental results show that our approach significantly activates the spatial reasoning ability of MLLM, improving performance on both in-domain and out-of-domain reasoning tasks. Our findings highlight the value of developing foundational spatial skills in MLLMs, supporting future progress in robotics, autonomous systems, and 3D scene understanding.

Finding the right north-star metrics is highly critical for advancing the mathematical reasoning capabilities of foundation models, especially given that existing evaluations are either too easy or only focus on getting correct short answers. To address these issues, we present IMO-Bench, a suite of advanced reasoning benchmarks, vetted by a panel of top specialists and that specifically targets the level of the International Mathematical Olympiad (IMO), the most prestigious venue for young mathematicians. IMO-AnswerBench first tests models on 400 diverse Olympiad problems with verifiable short answers. IMO-Proof Bench is the next-level evaluation for proof-writing capabilities, which includes both basic and advanced IMO level problems as well as detailed grading guidelines to facilitate automatic grading. These benchmarks played a crucial role in our historic achievement of the gold-level performance at IMO 2025 with Gemini Deep Think (Luong and Lockhart, 2025). Our model achieved 80.0% on IMO-AnswerBench and 65.7% on the advanced IMO-Proof Bench, surpassing the best non-Gemini models by large margins of 6.9% and 42.4% respectively. We also showed that autograders built with Gemini reasoning correlate well with human evaluations and construct IMO-GradingBench, with 1000 human gradings on proofs, to enable further progress in automatic evaluation of long-form answers. We hope that IMO-Bench will help the community towards advancing robust mathematical reasoning and release it at https://imobench.github.io/.

Foundation models in video generation are demonstrating remarkable capabilities as potential world models for simulating the physical world. However, their application in high-stakes domains like surgery, which demand deep, specialized causal knowledge rather than general physical rules, remains a critical unexplored gap. To systematically address this challenge, we present SurgVeo, the first expert-curated benchmark for video generation model evaluation in surgery, and the Surgical Plausibility Pyramid (SPP), a novel, four-tiered framework tailored to assess model outputs from basic appearance to complex surgical strategy. On the basis of the SurgVeo benchmark, we task the advanced Veo-3 model with a zero-shot prediction task on surgical clips from laparoscopic and neurosurgical procedures. A panel of four board-certified surgeons evaluates the generated videos according to the SPP. Our results reveal a distinct "plausibility gap": while Veo-3 achieves exceptional Visual Perceptual Plausibility, it fails critically at higher levels of the SPP, including Instrument Operation Plausibility, Environment Feedback Plausibility, and Surgical Intent Plausibility. This work provides the first quantitative evidence of the chasm between visually convincing mimicry and causal understanding in surgical AI. Our findings from SurgVeo and the SPP establish a crucial foundation and roadmap for developing future models capable of navigating the complexities of specialized, real-world healthcare domains.

Vision-language-action (VLA) models aim to understand natural language instructions and visual observations and to execute corresponding actions as an embodied agent. Recent work integrates future images into the understanding-acting loop, yielding unified VLAs that jointly understand, generate, and act -- reading text and images and producing future images and actions. However, these models either rely on external experts for modality unification or treat image generation and action prediction as separate processes, limiting the benefits of direct synergy between these tasks. Our core philosophy is to optimize generation and action jointly through a synchronous denoising process, where the iterative refinement enables actions to evolve from initialization, under constant and sufficient visual guidance. We ground this philosophy in our proposed Unified Diffusion VLA and Joint Discrete Denoising Diffusion Process (JD3P), which is a joint diffusion process that integrates multiple modalities into a single denoising trajectory to serve as the key mechanism enabling understanding, generation, and acting to be intrinsically synergistic. Our model and theory are built on a unified tokenized space of all modalities and a hybrid attention mechanism. We further propose a two-stage training pipeline and several inference-time techniques that optimize performance and efficiency. Our approach achieves state-of-the-art performance on benchmarks such as CALVIN, LIBERO, and SimplerEnv with 4times faster inference than autoregressive methods, and we demonstrate its effectiveness through in-depth analysis and real-world evaluations. Our project page is available at https://irpn-eai.github.io/UD-VLA.github.io/.

The remarkable success of multimodal large language models (MLLMs) has driven advances in multimodal embeddings, yet existing models remain inherently discriminative, limiting their ability to benefit from reasoning-driven generation paradigm. In this work, we pioneer the exploration of generative embeddings, unifying embedding tasks within a generative paradigm. We propose UME-R1, a universal multimodal embedding framework consisting of a two-stage training strategy: a cold-start supervised fine-tuning equips the model with reasoning capabilities and enables it to generate both discriminative and generative embeddings; a subsequent reinforcement learning enhances reasoning and further optimizes generative embedding quality. This pioneering work reveals four key insights: 1) generative embeddings unlock substantial performance gains over conventional discriminative embeddings by leveraging the powerful generative reasoning capabilities of MLLMs; 2) discriminative and generative embeddings are complementary, whose combined oracle performance far exceeding that of either alone; 3) RL can effectively enhance generative embeddings, establishing a scalable optimization paradigm.; 4) repeated sampling at inference boosts downstream task coverage (pass@k), highlighting the inference-time scalability potential of generative embeddings. Evaluated on the MMEB-V2 benchmark across 78 tasks spanning video, image, and visual documents, UME-R1 significantly outperforms conventional discriminative embedding models and offers a foundation for more interpretable, reasoning-driven generative multimodal embeddings. Our code, models, and datasets will be publicly available at https://github.com/XMUDeepLIT/UME-R1.

Graphical user interface (GUI) grounding is a key function of computer-use agents, which maps natural-language instructions to actionable screen regions. Existing approaches based on Multimodal Large Language Models (MLLMs) typically formulate it as a text-based coordinate generation task, yet directly generating precise coordinates from visual inputs remains challenging and computationally intensive. An intuitive way to implement GUI grounding is to first select visual patches relevant to the instructions and then determine the precise click location within those patches. Based on the observations that general MLLMs have some native grounding capability, nested within their attentions, we propose GUI-AIMA, an attention-based and coordinate-free supervised fine-tuning framework for efficient GUI grounding. GUI-AIMA aligns the intrinsic multimodal attention of MLLMs with patch-wise grounding signals. These signals are calculated adaptively for diverse user instructions by multi-head aggregation on simplified query-visual attention matrices. Besides, its coordinate-free manner can easily integrate a plug-and-play zoom-in stage. GUI-AIMA-3B was trained with only 85k screenshots, demonstrating exceptional data efficiency and verifying that light training can trigger the native grounding capability of MLLMs. It achieves state-of-the-art performance among 3B models, attaining an average accuracy of 58.6% on ScreenSpot-Pro and 62.2% on OSWorld-G. Project page: https://github.com/sjz5202/GUI-AIMA

Large Language Models (LLMs) have demonstrated strong capabilities in natural language reasoning, yet their application to Cyber Threat Intelligence (CTI) remains limited. CTI analysis involves distilling large volumes of unstructured reports into actionable knowledge, a process where LLMs could substantially reduce analyst workload. CTIBench introduced a comprehensive benchmark for evaluating LLMs across multiple CTI tasks. In this work, we extend CTIBench by developing AthenaBench, an enhanced benchmark that includes an improved dataset creation pipeline, duplicate removal, refined evaluation metrics, and a new task focused on risk mitigation strategies. We evaluate twelve LLMs, including state-of-the-art proprietary models such as GPT-5 and Gemini-2.5 Pro, alongside seven open-source models from the LLaMA and Qwen families. While proprietary LLMs achieve stronger results overall, their performance remains subpar on reasoning-intensive tasks, such as threat actor attribution and risk mitigation, with open-source models trailing even further behind. These findings highlight fundamental limitations in the reasoning capabilities of current LLMs and underscore the need for models explicitly tailored to CTI workflows and automation.

Natural Language Explanations (NLEs) describe how Large Language Models (LLMs) make decisions, drawing on both external Context Knowledge (CK) and Parametric Knowledge (PK) stored in model weights. Understanding their interaction is key to assessing the grounding of NLEs, yet it remains underexplored. Prior work has largely examined only single-step generation, typically the final answer, and has modelled PK and CK interaction only as a binary choice in a rank-1 subspace. This overlooks richer forms of interaction, such as complementary or supportive knowledge. We propose a novel rank-2 projection subspace that disentangles PK and CK contributions more accurately and use it for the first multi-step analysis of knowledge interactions across longer NLE sequences. Experiments on four QA datasets and three open-weight instruction-tuned LLMs show that diverse knowledge interactions are poorly represented in a rank-1 subspace but are effectively captured in our rank-2 formulation. Our multi-step analysis reveals that hallucinated NLEs align strongly with the PK direction, context-faithful ones balance PK and CK, and Chain-of-Thought prompting for NLEs shifts generated NLEs toward CK by reducing PK reliance. This work provides the first framework for systematic studies of multi-step knowledge interactions in LLMs through a richer rank-2 subspace disentanglement. Code and data: https://github.com/copenlu/pk-ck-knowledge-disentanglement.

In the retrieval domain, candidates' fusion from heterogeneous retrievers is a long-standing challenge, particularly for complex, multi-modal data such as videos. While typical fusion techniques are training-free, they rely solely on rank or score signals, disregarding candidates' representations. This work introduces Vote-in-Context (ViC), a generalized, training-free framework that re-thinks list-wise reranking and fusion as a zero-shot reasoning task for a Vision-Language Model (VLM). The core insight is to serialize both content evidence and retriever metadata directly within the VLM's prompt, allowing the model to adaptively weigh retriever consensus against visual-linguistic content. We demonstrate the generality of this framework by applying it to the challenging domain of cross-modal video retrieval. To this end, we introduce the S-Grid, a compact serialization map that represents each video as an image grid, optionally paired with subtitles to enable list-wise reasoning over video candidates. ViC is evaluated both as a single-list reranker, where it dramatically improves the precision of individual retrievers, and as an ensemble fuser, where it consistently outperforms strong baselines like CombSUM. Across video retrieval benchmarks including ActivityNet and VATEX, the framework establishes new state-of-the-art zero-shot retrieval performance, demonstrating its effectiveness in handling complex visual and temporal signals alongside text. In zero-shot settings, ViC achieves Recall@1 scores of 87.1% (t2v) / 89.0% (v2t) on MSR-VTT and 99.6% (v2t) on VATEX, representing massive gains of up to +40 Recall@1 over previous state-of-the-art baselines. We present ViC as a simple, reproducible, and highly effective recipe for turning modern VLMs into powerful zero-shot rerankers and fusers. Code and resources are publicly available at: https://github.com/mohammad2012191/ViC