| 2026-11-20 | 2611.26268 | Accelerating Mixture of Block Attention via Dynamic Routing, KV-Centric Execution and Runtime Pipelining | Yingqi Shi et al. | arXiv | 12 | |
| 2026-11-19 | 2611.26409 | WeMM-Video: Pushing the Frontiers of Open-Source Omni-Modal Video Understanding | Changli Tang et al. | arXiv | 40 | |
| 2026-11-19 | 2611.23824 | Evidence-Based Credit Assignment for Search-Augmented Language Models | Wen Wang et al. | arXiv | 6 | |
| 2026-11-16 | 2611.21373 | WeVisForensics: Self-Evolving Agent for Omni AI-Generated Video Detection | Yipeng Du et al. | arXiv | 49 | |
| 2026-11-05 | 2611.03404 | VidDub: Replacement-Aware Distillation for Two-Step Video Dubbing | Yexin Liu et al. | arXiv | 8 | |
| 2026-11-03 | 2611.04814 | Think Less, Locally: Local Preference Learning for Reasoning Compression | Muyao Yuan et al. | arXiv | 56 | |
| 2026-11-02 | 2610.40339 | Wizard: Manipulate Any World in Real Time via Agentic Control | Yuzhe Wu et al. | arXiv | 27 | |
| 2026-11-02 | 2610.16927 | WeFair: Fidelity-Aware Alignment from Human Feedback for Personalized Query Rewriting in Short-Video Search | Yingtao Ren et al. | arXiv | 23 | |
| 2026-10-29 | 2610.39301 | Youtu-Parsing-Omni: One Encoder, One Schema, All Modalities | Yunhang Shen et al. | arXiv | 13 |  |
| 2026-10-28 | 2610.38352 | WeVisCut: The Whole Editing Room in One Agent From Narration to Long-Form Video | Chenyu Mu et al. | arXiv | 15 | |
| 2026-10-23 | 2610.32411 | A Broader Look into MLLM's Retrieval Heads | Jia Wang et al. | arXiv | 28 | |
| 2026-10-22 | 2610.31162 | HiSTEVE 1.0: Harnessing Human-Interface VLA Agents for Open-World Games | Zecong Tang et al. | arXiv | 58 | |
| 2026-10-22 | 2610.30926 | Making MLLM Latents Matter: Factor-Aware Predictive Learning for Cinematic Video Generation | Songlin Yang et al. | arXiv | 50 | |
| 2026-10-22 | 2610.31329 | WeVisOne-VL: Unified Visual Perception from Objects to Pixels | Tianming Liang et al. | arXiv | 48 | |
| 2026-10-21 | 2610.28069 | Training the Whole Agent, Not Just the Model | Lingxiao Tang et al. | arXiv | 19 | |
| 2026-10-21 | 2610.19425 | Evaluating Agentic Capabilities Before Post-Training: An End-to-End Approach | Jiarui Qin et al. | arXiv | 10 | |
| 2026-10-16 | 2610.20693 | Orbia: Do Generated Videos Form a Consistent 3D World? | Dongyue Lu et al. | arXiv | 48 | |
| 2026-10-16 | 2610.22192 | Terminal-RSI: Recursive Self-Improving Terminal Agents via Recursive Self-Rewrite and Policy-Refine | Junyao Yang et al. | arXiv | 37 | |
| 2026-10-15 | 2610.17242 | When Left Is Right? Camera-Motion Understanding beyond Screen | Chong Cheng et al. | arXiv | 6 | |
| 2026-10-15 | 2610.16338 | SyncReward: Human-Aligned Synchronization Reward for Audio-Visual Generation | Kaisi Guan et al. | arXiv | 4 | |
| 2026-10-14 | 2610.15765 | A Decade in a Day: Synthesizing Agent-Native Interfaces from Human-Oriented Software Evolution | Alan Wu et al. | arXiv | 29 | |
| 2026-10-05 | 2609.37925 | Rollout-Marginal Distillation for Long-Horizon Autoregressive Video Generation | Chenjian Gao et al. | arXiv | 8 | |
| 2026-10-02 | 2609.37200 | Adaptive Reward Routing: Dynamic Multi-Reward Optimization for Joint Audio-Video Diffusion via Forward-Process RL | Songlin Yang et al. | arXiv | 22 | |
| 2026-09-29 | 2609.34972 | Just MLPs: Efficient Visual State Reconstruction for Multimodal Language Models | Jingdi lei et al. | arXiv | 27 | |
| 2026-09-29 | 2609.32472 | AdaTutoRank: Learning to Rerank Document Sets via Adaptive Tutoring Optimization for RAG and Deep Research | Kailin Jiang et al. | arXiv | 19 | |
| 2026-09-29 | 2609.34754 | Draft-KV: Learning Useful Latent Communication Between Language Models | Linquan Wu et al. | arXiv | 8 | |
| 2026-09-28 | 2609.29050 | SLCA-GRPO: Resolving Cross-Segment Credit Misattribution in Tool-Calling RL | Yan Zhan et al. | arXiv | 11 | |
| 2026-09-24 | 2609.22947 | RewardVerse: Rubric-Guided Policy Optimization for Video Reward Modeling | Zhenchen Tang et al. | arXiv | 36 | |
| 2026-09-22 | 2609.25001 | GameHorizon Suite: Multi-Horizon Data and Evaluation in Gameplay | Yiran Wang et al. | arXiv | 52 | |
| 2026-09-21 | 2609.22069 | OmniVBench: A Benchmark and Large-Scale Dataset for Omni Reference-to-Video Generation | Wenxue Li et al. | arXiv | 33 | |
| 2026-09-18 | 2609.20423 | WeVisDoc: From Coverage to Capability for Robust End-to-End Document Parsing | Hao Yu et al. | arXiv | 24 | |
| 2026-08-31 | 2608.28476 | ContextPilot: Teaching Agents for Proactive Context Management via Fine-grained RL | Zhuoshi Pan et al. | arXiv | 31 | |
| 2026-08-31 | 2608.28478 | Blind Men and the Elephant: Probing the Epistemic Myopia of LLMs under Long-Tail Divergent Knowledge | Zhuoshi Pan et al. | arXiv | 19 | |
| 2026-08-28 | 2608.26200 | GameWAM: A World Action Model for Video Games | Yuncheng Guo et al. | arXiv | 35 | |
| 2026-08-26 | 2608.24053 | WeMM-Embedding: WeChat Multi-Modal Embedding Technical Report | Junjie Zhou et al. | arXiv | 73 | |
| 2026-08-24 | 2608.12781 | Beyond Correctness: Benchmarking and Aligning Response Behaviors in Hybrid-Thinking MLLMs | Xinming Wang et al. | arXiv | 35 | |
| 2026-08-21 | 2608.13120 | SkillEvo: Self-Renewing Evolution Gradients from Multi-Turn Interaction Feedback | Qianxi Yan et al. | arXiv | 26 | |
| 2026-08-21 | 2608.19758 | FlashPrefill V2: Block-Sparse Prefill Attention for Long-Context LLM Serving | Qihang Fan et al. | arXiv | 16 | |
| 2026-08-19 | 2608.17566 | CoinVE-200K: A Large-Scale High-Quality Dataset for Compositional Instruction-Guided Video Editing | Fuchen Long et al. | arXiv | 16 | |
| 2026-08-19 | 2608.16393 | Security Assessment of DeepSeek Harness with A.I.G: Evaluating Resistance to Indirect Prompt Injection | Zonghao Ying et al. | arXiv | 7 | |
| 2026-08-18 | 2608.15265 | VibeWorlding: Can Multimodal Agents Construct 3D Open Worlds End-to-End? | Yansong Ning et al. | arXiv | 62 | |
| 2026-08-18 | 2608.16391 | Ventor-QTest: Threat-Model-Driven Verification of Vendor-Hosted LLM APIs | Xiangfan Wu et al. | arXiv | 13 | |
| 2026-08-10 | 2608.07051 | YOLO-PEFT: Parameter-Efficient Fine-Tuning on YOLO Family | Xu Lin et al. | arXiv | 21 | |
| 2026-08-07 | 2608.06197 | EnvACE: Internalizing Environment Dynamics via World Rehearsal for Agentic Reinforcement Learning | Zishan Xu et al. | arXiv | 39 | |
| 2026-08-06 | 2608.04574 | When Memory Lies: An Empirical Study of Spatial Memory Staleness in VLM Agents | Yushi Sun et al. | arXiv | 11 | |
| 2026-08-06 | 2608.04964 | WorldCycle: Self-Verifiable Reinforcement Learning for Long-Horizon Video World Models | Bohai Gu et al. | arXiv | 10 | |
| 2026-08-05 | 2608.03509 | SkillJack: Persistent Skill Backdoors in Self-Evolving Agents | Zonghao Ying et al. | arXiv | 11 | |
| 2026-07-29 | 2607.24223 | A New Role for Relevance: Guiding Corpus Interaction in Agentic Search | Jiangnan Li et al. | arXiv | 84 | |
| 2026-07-24 | 2607.20061 | ReferTrack: Referring Then Tracking for Embodied Visual Tracking | Hanjing Ye et al. | arXiv | 46 | |
| 2026-07-24 | 2607.20911 | Tencent WorkBuddy Bench: A Multi-Domain Coding-Agent Benchmark with Contamination-Resistant Task Construction | Tencent WorkBuddy Bench Team et al. | arXiv | 22 | |
| 2026-07-23 | 2607.19747 | Beyond Relevance-Centric Retrieval: Rubric-Oriented Document Set Selection and Ranking | Kailin Jiang et al. | arXiv | 28 | |
| 2026-07-21 | 2607.17250 | EvolvingWorld: An Open-Schema Framework for Co-Evolving Role-Play Agents and World Model in Interactive Literary World | Qing Zong et al. | arXiv | 87 | |
| 2026-07-08 | 2607.04884 | HunyuanOCR-1.5: Making Lightweight OCR VLMs Faster and Better | Gengluo Li et al. | arXiv | 10 | |
| 2026-07-08 | 2607.00394 | When Classic Cache Policies Fail: Learning-Augmented Replacement for Semantic Retrieval Buffers | Yushi Sun et al. | arXiv | 4 | |
| 2026-07-07 | 2607.03748 | Bridging Interleaved Multi-Modal Reasoning as a Unified Decision Process | Zican Hu et al. | arXiv | 39 | |
| 2026-07-06 | 2606.31227 | Securing the AI Agent: A Unified Framework for Multi-Layer Agent Red Teaming | Yong Yang et al. | arXiv | 12 | |
| 2026-07-01 | 2606.08671 | SkillHone: A Harness for Continual Agent Skill Evolution Through Persistent Decision History | Zhiwei Li et al. | arXiv | 17 | |
| 2026-06-30 | 2502.16886 | ReFreeKV: Towards Threshold-Free KV Cache Compression | Xuanfan Ni et al. | arXiv | 42 | |
| 2026-06-23 | 2606.11075 | Exploring the Design Space of Reward Backpropagation for Flow Matching | Ruoyu Wang et al. | arXiv | 11 | |
| 2026-06-15 | 2606.14502 | From Chatbot to Digital Colleague: The Paradigm Shift Toward Persistent Autonomous AI | Yongheng Zhang et al. | arXiv | 62 | |
| 2026-06-11 | 2606.11119 | TRACE: A Unified Rollout Budget Allocation Framework for Efficient Agentic Reinforcement Learning | Heming Zou et al. | arXiv | 21 | |
| 2026-06-09 | 2606.09079 | FlashMemory-DeepSeek-V4: Lightning Index Ultra-Long Context via Lookahead Sparse Attention | Yan Wang et al. | arXiv | 68 | |
| 2026-06-09 | 2605.25077 | WorldCraft: From Camera Navigation to Object Manipulation in Interactive Video World Models | Bohai Gu et al. | arXiv | 21 | |
| 2026-06-03 | 2606.03603 | World Models Meet Language Models: On the Complementarity of Concrete and Abstract Reasoning | Yucheng Zhou et al. | arXiv | 27 | |
| 2026-05-28 | 2605.28398 | HRBench: Benchmarking and Understanding Thinking-Mode Switch Strategies in Hybrid-Reasoning LLMs | Yansong Ning et al. | arXiv | 13 | |
| 2026-05-27 | 2605.23271 | EvalVerse: Pipeline-Aware and Expert-Calibrated Benchmarking for Professional Cinematic Video Generation | Songlin Yang et al. | arXiv | 30 | |
| 2026-05-27 | 2605.26952 | Efficient Agentic Reinforcement Learning with On-Policy Intrinsic Knowledge Boundary Enhancement | Dingwei Chen et al. | arXiv | 11 | |
| 2026-05-26 | 2605.25343 | Toward Native Multimodal Modeling: A Roadmap | Siyu An et al. | arXiv | 38 | |
| 2026-05-21 | 2605.19597 | HyperLogic: A Hard, Forward-Authored Chinese Logical Reasoning Benchmark with Execution-Derived Answers | Ming Zhang et al. | arXiv | 19 | |
| 2026-05-11 | 2605.06221 | UniPrefill: Universal Long-Context Prefill Acceleration via Block-wise Dynamic Sparsification | Qihang Fan et al. | arXiv | 20 | |
| 2026-05-08 | 2605.06416 | MiA-Signature: Approximating Global Activation for Long-Context Understanding | Yuqing Li et al. | arXiv | 54 | |
| 2026-05-08 | 2605.06200 | A^2TGPO: Agentic Turn-Group Policy Optimization with Adaptive Turn-level Clipping | Dingwei Chen et al. | arXiv | 13 | |