AgentMath: Empowering Mathematical Reasoning for Large Language Models via Tool-Augmented Agent

Large Reasoning Models (LRMs) like o3 and DeepSeek-R1 have achieved remarkable progress in natural language reasoning with long chain-of-thought. However, they remain computationally inefficient and struggle with accuracy when solving problems requiring complex mathematical operations. In this work, we present AgentMath, an agent framework that seamlessly integrates language models' reasoning capabilities with code interpreters' computational precision to efficiently tackle complex mathematical problems. Our approach introduces three key innovations: (1) An automated method that converts natural language chain-of-thought into structured tool-augmented trajectories, generating high-quality supervised fine-tuning (SFT) data to alleviate data scarcity; (2) A novel agentic reinforcement learning (RL) paradigm that dynamically interleaves natural language generation with real-time code execution. This enables models to autonomously learn optimal tool-use strategies through multi-round interactive feedback, while fostering emergent capabilities in code refinement and error correction; (3) An efficient training system incorporating innovative techniques, including request-level asynchronous rollout scheduling, agentic partial rollout, and prefix-aware weighted load balancing, achieving 4-5x speedup and making efficient RL training feasible on ultra-long sequences with scenarios with massive tool calls.Extensive evaluations show that AgentMath achieves state-of-the-art performance on challenging mathematical competition benchmarks including AIME24, AIME25, and HMMT25. Specifically, AgentMath-30B-A3B attains 90.6%, 86.4%, and 73.8% accuracy respectively, achieving advanced capabilities.These results validate the effectiveness of our approach and pave the way for building more sophisticated and scalable mathematical reasoning agents.

RealisticDreamer: Guidance Score Distillation for Few-shot Gaussian Splatting

3D Gaussian Splatting (3DGS) has recently gained great attention in the 3D scene representation for its high-quality real-time rendering capabilities. However, when the input comprises sparse training views, 3DGS is prone to overfitting, primarily due to the lack of intermediate-view supervision. Inspired by the recent success of Video Diffusion Models (VDM), we propose a framework called Guidance Score Distillation (GSD) to extract the rich multi-view consistency priors from pretrained VDMs. Building on the insights from Score Distillation Sampling (SDS), GSD supervises rendered images from multiple neighboring views, guiding the Gaussian splatting representation towards the generative direction of VDM. However, the generative direction often involves object motion and random camera trajectories, making it challenging for direct supervision in the optimization process. To address this problem, we introduce an unified guidance form to correct the noise prediction result of VDM. Specifically, we incorporate both a depth warp guidance based on real depth maps and a guidance based on semantic image features, ensuring that the score update direction from VDM aligns with the correct camera pose and accurate geometry. Experimental results show that our method outperforms existing approaches across multiple datasets.

Force-Modulated Visual Policy for Robot-Assisted Dressing with Arm Motions

Robot-assisted dressing has the potential to significantly improve the lives of individuals with mobility impairments. To ensure an effective and comfortable dressing experience, the robot must be able to handle challenging deformable garments, apply appropriate forces, and adapt to limb movements throughout the dressing process. Prior work often makes simplifying assumptions -- such as static human limbs during dressing -- which limits real-world applicability. In this work, we develop a robot-assisted dressing system capable of handling partial observations with visual occlusions, as well as robustly adapting to arm motions during the dressing process. Given a policy trained in simulation with partial observations, we propose a method to fine-tune it in the real world using a small amount of data and multi-modal feedback from vision and force sensing, to further improve the policy's adaptability to arm motions and enhance safety. We evaluate our method in simulation with simplified articulated human meshes and in a real world human study with 12 participants across 264 dressing trials. Our policy successfully dresses two long-sleeve everyday garments onto the participants while being adaptive to various kinds of arm motions, and greatly outperforms prior baselines in terms of task completion and user feedback. Video are available at https://dressing-motion.github.io/.

Geometric Red-Teaming for Robotic Manipulation

Standard evaluation protocols in robotic manipulation typically assess policy performance over curated, in-distribution test sets, offering limited insight into how systems fail under plausible variation. We introduce Geometric Red-Teaming (GRT), a red-teaming framework that probes robustness through object-centric geometric perturbations, automatically generating CrashShapes -- structurally valid, user-constrained mesh deformations that trigger catastrophic failures in pre-trained manipulation policies. The method integrates a Jacobian field-based deformation model with a gradient-free, simulator-in-the-loop optimization strategy. Across insertion, articulation, and grasping tasks, GRT consistently discovers deformations that collapse policy performance, revealing brittle failure modes missed by static benchmarks. By combining task-level policy rollouts with constraint-aware shape exploration, we aim to build a general purpose framework for structured, object-centric robustness evaluation in robotic manipulation. We additionally show that fine-tuning on individual CrashShapes, a process we refer to as blue-teaming, improves task success by up to 60 percentage points on those shapes, while preserving performance on the original object, demonstrating the utility of red-teamed geometries for targeted policy refinement. Finally, we validate both red-teaming and blue-teaming results with a real robotic arm, observing that simulated CrashShapes reduce task success from 90% to as low as 22.5%, and that blue-teaming recovers performance to up to 90% on the corresponding real-world geometry -- closely matching simulation outcomes. Videos and code can be found on our project website: https://georedteam.github.io/ .

See What You Need: Query-Aware Visual Intelligence through Reasoning-Perception Loops

Human video comprehension demonstrates dynamic coordination between reasoning and visual attention, adaptively focusing on query-relevant details. However, current long-form video question answering systems employ rigid pipelines that decouple reasoning from perception, leading to either information loss through premature visual abstraction or computational inefficiency through exhaustive processing. The core limitation lies in the inability to adapt visual extraction to specific reasoning requirements, different queries demand fundamentally different visual evidence from the same video content. In this work, we present CAVIA, a training-free framework that revolutionizes video understanding through reasoning, perception coordination. Unlike conventional approaches where visual processing operates independently of reasoning, CAVIA creates a closed-loop system where reasoning continuously guides visual extraction based on identified information gaps. CAVIA introduces three innovations: (1) hierarchical reasoning, guided localization to precise frames; (2) cross-modal semantic bridging for targeted extraction; (3) confidence-driven iterative synthesis. CAVIA achieves state-of-the-art performance on challenging benchmarks: EgoSchema (65.7%, +5.3%), NExT-QA (76.1%, +2.6%), and IntentQA (73.8%, +6.9%), demonstrating that dynamic reasoning-perception coordination provides a scalable paradigm for video understanding.

Demystifying the Visual Quality Paradox in Multimodal Large Language Models

Recent Multimodal Large Language Models (MLLMs) excel on benchmark vision-language tasks, yet little is known about how input visual quality shapes their responses. Does higher perceptual quality of images already translate to better MLLM understanding? We conduct the first systematic study spanning leading MLLMs and a suite of vision-language benchmarks, applying controlled degradations and stylistic shifts to each image. Surprisingly, we uncover a visual-quality paradox: model, task, and even individual-instance performance can improve when images deviate from human-perceived fidelity. Off-the-shelf restoration pipelines fail to reconcile these idiosyncratic preferences. To close the gap, we introduce Visual-Quality Test-Time Tuning (VQ-TTT)-a lightweight adaptation module that: (1) inserts a learnable, low-rank kernel before the frozen vision encoder to modulate frequency content; and (2) fine-tunes only shallow vision-encoder layers via LoRA. VQ-TTT dynamically adjusts each input image in a single forward pass, aligning it with task-specific model preferences. Across the evaluated MLLMs and all datasets, VQ-TTT lifts significant average accuracy, with no external models, cached features, or extra training data. These findings redefine ``better'' visual inputs for MLLMs and highlight the need for adaptive, rather than universally ``clean'', imagery, in the new era of AI being the main data customer.

Stepwise Reasoning Checkpoint Analysis: A Test Time Scaling Method to Enhance LLMs' Reasoning

Mathematical reasoning through Chain-of-Thought (CoT) has emerged as a powerful capability of Large Language Models (LLMs), which can be further enhanced through Test-Time Scaling (TTS) methods like Beam Search and DVTS. However, these methods, despite improving accuracy by allocating more computational resources during inference, often suffer from path homogenization and inefficient use of intermediate results. To address these limitations, we propose Stepwise Reasoning Checkpoint Analysis (SRCA), a framework that introduces checkpoints between reasoning steps. It incorporates two key strategies: (1) Answer-Clustered Search, which groups reasoning paths by their intermediate checkpoint answers to maintain diversity while ensuring quality, and (2) Checkpoint Candidate Augmentation, which leverages all intermediate answers for final decision-making. Our approach effectively reduces path homogenization and creates a fault-tolerant mechanism by utilizing high-quality intermediate results. Experimental results show that SRCA improves reasoning accuracy compared to existing TTS methods across various mathematical datasets.

Sparc3D: Sparse Representation and Construction for High-Resolution 3D Shapes Modeling

High-fidelity 3D object synthesis remains significantly more challenging than 2D image generation due to the unstructured nature of mesh data and the cubic complexity of dense volumetric grids. Existing two-stage pipelines-compressing meshes with a VAE (using either 2D or 3D supervision), followed by latent diffusion sampling-often suffer from severe detail loss caused by inefficient representations and modality mismatches introduced in VAE. We introduce Sparc3D, a unified framework that combines a sparse deformable marching cubes representation Sparcubes with a novel encoder Sparconv-VAE. Sparcubes converts raw meshes into high-resolution ($1024^3$) surfaces with arbitrary topology by scattering signed distance and deformation fields onto a sparse cube, allowing differentiable optimization. Sparconv-VAE is the first modality-consistent variational autoencoder built entirely upon sparse convolutional networks, enabling efficient and near-lossless 3D reconstruction suitable for high-resolution generative modeling through latent diffusion. Sparc3D achieves state-of-the-art reconstruction fidelity on challenging inputs, including open surfaces, disconnected components, and intricate geometry. It preserves fine-grained shape details, reduces training and inference cost, and integrates naturally with latent diffusion models for scalable, high-resolution 3D generation.

Rethinking Stateful Tool Use in Multi-Turn Dialogues: Benchmarks and Challenges

Existing benchmarks that assess Language Models (LMs) as Language Agents (LAs) for tool use primarily focus on stateless, single-turn interactions or partial evaluations, such as tool selection in a single turn, overlooking the inherent stateful nature of interactions in multi-turn applications. To fulfill this gap, we propose \texttt{DialogTool}, a multi-turn dialogue dataset with stateful tool interactions considering the whole life cycle of tool use, across six key tasks in three stages: 1) \textit{tool creation}; 2) \textit{tool utilization}: tool awareness, tool selection, tool execution; and 3) \textit{role-consistent response}: response generation and role play. Furthermore, we build \texttt{VirtualMobile} -- an embodied virtual mobile evaluation environment to simulate API calls and assess the robustness of the created APIs\footnote{We will use tools and APIs alternatively, there are no significant differences between them in this paper.}. Taking advantage of these artifacts, we conduct comprehensive evaluation on 13 distinct open- and closed-source LLMs and provide detailed analysis at each stage, revealing that the existing state-of-the-art LLMs still cannot perform well to use tools over long horizons.

Pangu Ultra MoE: How to Train Your Big MoE on Ascend NPUs

Sparse large language models (LLMs) with Mixture of Experts (MoE) and close to a trillion parameters are dominating the realm of most capable language models. However, the massive model scale poses significant challenges for the underlying software and hardware systems. In this paper, we aim to uncover a recipe to harness such scale on Ascend NPUs. The key goals are better usage of the computing resources under the dynamic sparse model structures and materializing the expected performance gain on the actual hardware. To select model configurations suitable for Ascend NPUs without repeatedly running the expensive experiments, we leverage simulation to compare the trade-off of various model hyperparameters. This study led to Pangu Ultra MoE, a sparse LLM with 718 billion parameters, and we conducted experiments on the model to verify the simulation results. On the system side, we dig into Expert Parallelism to optimize the communication between NPU devices to reduce the synchronization overhead. We also optimize the memory efficiency within the devices to further reduce the parameter and activation management overhead. In the end, we achieve an MFU of 30.0% when training Pangu Ultra MoE, with performance comparable to that of DeepSeek R1, on 6K Ascend NPUs, and demonstrate that the Ascend system is capable of harnessing all the training stages of the state-of-the-art language models. Extensive experiments indicate that our recipe can lead to efficient training of large-scale sparse language models with MoE. We also study the behaviors of such models for future reference.