| SkillGrad: Optimizing Agent Skills Like Gradient Descent | Agents | Hanyu Wang +4 | May 26, 2026 | 27 |
| MUSE-Autoskill: Self-Evolving Agents via Skill Creation, Memory, Management, and Evaluation | Agents | Huawei Lin +4 | May 26, 2026 | 27 |
| LLaVA-OneVision-2: Towards Next-Generation Perceptual Intelligence | Robotics | Xiang An +29 | May 25, 2026 | 27 |
| EVA-Client: A Unified Data Collection, Inference, and Deployment Framework for Embodied Policies on Real Robots | Robotics | Heqing Yang +10 | Jul 2, 2026 | 26 |
| Reinforcement Learning with Metacognitive Feedback Elicits Faithful Uncertainty Expression in LLMs | Reinforcement learning | Gabrielle Kaili-May Liu +4 | Jun 30, 2026 | 26 |
| From Trainee to Trainer: LLM-Designed Training Environment for RL with Multi-Agent Reasoning | Agents | Chao Chen +4 | Jun 16, 2026 | 26 |
| Smaller Models are Natural Explorers for Policy-Level Diversity in GRPO | Reinforcement learning | Yiming Ren +10 | Jun 2, 2026 | 26 |
| World Pilot: Steering Vision-Language-Action Models with World-Action Priors | Robotics | Zefu Lin +6 | Jun 10, 2026 | 26 |
| Direct 3D-Aware Object Insertion via Decomposed Visual Proxies | Robotics | Jingbo Gong +8 | Jun 4, 2026 | 26 |
| RobotValues: Evaluating Household Robots When Human Values Conflict | Robotics | Jongwook Han +2 | Jun 2, 2026 | 26 |
| M^3Eval: Multi-Modal Memory Evaluation through Cognitively-Grounded Video Tasks | Vision & multimodal | Jie Huang +6 | Jun 3, 2026 | 26 |
| AutoMedBench: Towards Medical AutoResearch with Agentic AI Models | Vision & multimodal | Junqi Liu +14 | Jun 1, 2026 | 26 |
| VideoMLA: Low-Rank Latent KV Cache for Minute-Scale Autoregressive Video Diffusion | Image & video gen | Hidir Yesiltepe +6 | May 28, 2026 | 26 |
| VLM3: Vision Language Models Are Native 3D Learners | Vision & multimodal | Zhipeng Cai +5 | May 28, 2026 | 26 |
| Not All Disagreement Is Learnable: Token Teachability in On-Policy Distillation | Reinforcement learning | Yuanyi Wang +8 | May 26, 2026 | 26 |
| UniSteer: Text-Guided Flow Matching in Activation Space for Versatile LLM Steering | Efficiency & systems | Yingdong Shi +6 | May 28, 2026 | 26 |
| AI Research Agents Narrow Scientific Exploration | Agents | Yixuan Tang +1 | May 27, 2026 | 26 |
| Parallelized Autoregressive Decoding for Omni-Modal Dense Video Captioning | Speech & audio | Wenzheng Zeng +4 | Jul 3, 2026 | 25 |
| InternVLA-A1.5: Unifying Understanding, Latent Foresight, and Action for Compositional Generalization | Robotics | Haoxiang Ma +28 | Jul 6, 2026 | 25 |
| Video-MME-Logical: A Controlled Diagnostic Benchmark for Video Temporal-Logical Reasoning | Vision & multimodal | Hohin Kwan +7 | Jun 26, 2026 | 25 |
| Causal-rCM: A Unified Teacher-Forcing and Self-Forcing Open Recipe for Autoregressive Diffusion Distillation in Streaming Video Generation and Interactive World Models | Image & video gen | Kaiwen Zheng +9 | Jun 24, 2026 | 25 |
| End-to-End Context Compression at Scale | Agents | Ang Li +14 | Jun 8, 2026 | 25 |
| Reinforcement Learning Elicits Contextual Learning of Unseen Language Translation | Code | Hanxu Hu +4 | Jun 4, 2026 | 25 |
| ThoughtFold: Folding Reasoning Chains via Introspective Preference Learning | Reinforcement learning | Ziyan Liu +9 | Jun 2, 2026 | 25 |
| Benchmarks are Not Enough: RAMP for Runtime Assessing of Agentic Models in Production Systems | Agents | Yipeng Ouyang +5 | May 26, 2026 | 25 |
| Which Pretraining Paradigm Better Serves Spatial Intelligence? An Empirical Comparison of Vision-Language and Video Generation Models | Image & video gen | Haozhan Shen +3 | May 27, 2026 | 25 |
| SAAS: Self-Aware Reinforcement Learning for Over-Search Mitigation in Agentic Search | Agents | Yunbo Tang +6 | May 28, 2026 | 25 |
| Colored Noise Diffusion Sampling | Image & video gen | Hadar Davidson +2 | May 28, 2026 | 25 |
| Triplet-Block Diffusion RWKV | Image & video gen | Ke Lin +4 | May 25, 2026 | 25 |
| Managing Procedural Memory in LLM Agents: Control, Adaptation, and Evaluation | Agents | Julia Belikova +6 | Jun 22, 2026 | 24 |
| One-Step Gradient Delay is Not a Barrier for Large-Scale Asynchronous Pipeline Parallel LLM Pretraining | Efficiency & systems | Philip Zmushko +4 | Jun 29, 2026 | 24 |
| BioMatrix: Towards a Comprehensive Biological Foundation Model Spanning the Modality Matrix of Sequences, Structures, and Language | Vision & multimodal | Qizhi Pei +11 | Jun 20, 2026 | 24 |
| Deeper is Not Always Better: Mitigating the Alignment Tax via Confident Layer Decoding | Code | Xuanming Zhang +10 | Jun 20, 2026 | 24 |
| EfficientRollout: System-Aware Self-Speculative Decoding for RL Rollouts | Agents | Minseo Kim +9 | Jun 17, 2026 | 24 |
| N-GRPO: Embedding-Level Neighbor Mixing for Enhanced Policy Optimization | Reinforcement learning | Xukun Zhu +3 | Jun 9, 2026 | 24 |
| ARM: An AutoRegressive Large Multimodal Model with Unified Discrete Representations | Image & video gen | Junke Wang +18 | Jun 9, 2026 | 24 |
| Skill-3D: Evolving Scene-Aware Skills for Agentic 3D Spatial Reasoning | Vision & multimodal | Haoyuan Li +4 | Jun 5, 2026 | 24 |
| LoomVideo: Unifying Multimodal Inputs into Video Generation and Editing | Image & video gen | Jianzong Wu +13 | Jun 4, 2026 | 24 |
| Self-Distilled Policy Gradient | Code | Yifeng Liu +3 | Jun 2, 2026 | 24 |
| When Should Models Change Their Minds? Contextual Belief Management in Large Language Models | Code | Haoming Xu +8 | May 28, 2026 | 24 |
| OSP-Next: Efficient High-Quality Video Generation with Sparse Sequence Parallelism, HiF8 Quantization, and Reinforcement Learning | Image & video gen | Yunyang Ge +6 | May 27, 2026 | 24 |
| QUACK: Questioning, Understanding, and Auditing Communicated Knowledge in Multimodal Social Deduction Agents | Vision & multimodal | Ye Yuan +14 | May 26, 2026 | 24 |
| AutoTrainess: Teaching Language Models to Improve Language Models Autonomously | Agents | Zhaojian Yu +5 | Jun 30, 2026 | 23 |
| FLAT: Feedforward Latent Triangle Splatting for Geometrically Accurate Scene Generation | Image & video gen | Orest Kupyn +5 | Jun 23, 2026 | 23 |
| LingxiDiagBench: A Multi-Agent Framework for Benchmarking LLMs in Chinese Psychiatric Consultation and Diagnosis | Agents | Shihao Xu +10 | Jun 11, 2026 | 23 |
| RedAct: Redacting Agent Capability Traces for Procedural Skill Protection | Vision & multimodal | Shuwen Xu +2 | Jun 10, 2026 | 23 |
| Skip a Layer or Loop It? Learning Program-of-Layers in LLMs | Efficiency & systems | Ziyue Li +2 | Jun 4, 2026 | 23 |
| On Subquadratic Architectures: From Applications to Principles | Code | Anamaria-Roberta Hartl +8 | Jun 10, 2026 | 23 |
| InternVideo3: Agentify Foundation Models with Multimodal Contextual Reasoning | Vision & multimodal | Ziang Yan +12 | Jun 10, 2026 | 23 |
| When Tools Fail: Benchmarking Dynamic Replanning and Anomaly Recovery in LLM Agents | Agents | Dongsheng Zhu +7 | Jun 4, 2026 | 23 |
| Benchmarking Visual State Tracking in Multimodal Video Understanding | Vision & multimodal | Sihyun Yu +10 | Jun 2, 2026 | 23 |
| WALL-WM: Carving World Action Modeling at the Event Joints | Robotics | Shalfun Li +30 | Jun 1, 2026 | 23 |
| Exploring Autonomous Agentic Data Engineering for Model Specialization | Agents | Yujie Luo +12 | May 28, 2026 | 23 |
| LoMo: Local Modality Substitution for Deeper Vision-Language Fusion | Vision & multimodal | Feng Han +4 | May 28, 2026 | 23 |
| GUI-CIDER: Mid-training GUI Agents via Causal Internalization and Density-aware Exemplar Reselection | Vision & multimodal | Zheng Wu +7 | May 27, 2026 | 23 |
| Long Live The Balance: Information Bottleneck Driven Tree-based Policy Optimization | Code | Hao Jiang +9 | May 27, 2026 | 23 |
| Gemini Embedding 2: A Native Multimodal Embedding Model from Gemini | Speech & audio | Madhuri Shanbhogue +39 | May 26, 2026 | 23 |
| Multiplayer Interactive World Models with Representation Autoencoders | Image & video gen | Anthony Hu +26 | Jul 6, 2026 | 22 |
| Domain Arithmetic: One-Shot VLA Adaptation under Environmental Shifts | Robotics | Taewook Kang +3 | Jul 1, 2026 | 22 |
| ASPIRE: Agentic /Skills Discovery for Robotics | Robotics | Runyu Lu +13 | Jun 30, 2026 | 22 |
| VideoSearch-R1: Iterative Video Retrieval and Reasoning via Soft Query Refinement | Vision & multimodal | Seohyun Lee +4 | Jul 1, 2026 | 22 |
| OSWorld2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks | Vision & multimodal | Mengqi Yuan +35 | Jun 28, 2026 | 22 |
| One Model, Many Latencies: Universal Speech Enhancement for Diverse Real-Time Applications | Speech & audio | Szu-Wei Fu +6 | Jun 24, 2026 | 22 |
| WorldCraft: From Camera Navigation to Object Manipulation in Interactive Video World Models | Robotics | Bohai Gu +11 | May 24, 2026 | 22 |
| NVIDIA OmniDreams: Real-Time Generative World Model for Closed-Loop Autonomous Vehicle Simulation | Robotics | NVIDIA +33 | Jun 2, 2026 | 22 |
| Bootstrap Your Generator: Unpaired Visual Editing with Flow Matching | Image & video gen | Yoad Tewel +3 | Jun 2, 2026 | 22 |
| MIRA: Mid-training Rubric Anchoring for Source-Aware Data Selection | Code | Haowen Wang +10 | May 28, 2026 | 22 |
| Light-Omni: Reflex over Reasoning in Agentic Video Understanding with Long-Term Memory | Vision & multimodal | Chang Nie +4 | Jul 6, 2026 | 21 |
| Do All Visual Tokens Matter Equally? Object-Evidence Preserving Token Merging for Vision-Language Retrieval | Vision & multimodal | Suhyeong Park +3 | Jul 6, 2026 | 21 |
| TACO: Tool-Augmented Credit Optimization for Agentic Tool Use | Vision & multimodal | Mingkuan Feng +7 | Jun 29, 2026 | 21 |
| Monte Carlo Energy Aggregation for Mobile 3D Gaussian Splatting | Vision & multimodal | Xiaobiao Du +5 | Jun 29, 2026 | 21 |
| Critique of Agent Model | Image & video gen | Eric Xing +2 | Jun 22, 2026 | 21 |
| MCompassRAG: Topic Metadata as a Semantic Compass for Paragraph-Level Retrieval | Code | Amirhossein Abaskohi +5 | Jun 16, 2026 | 21 |
| FlowBender: Feedback-Aware Training for Self-Correcting Conditional Flows | Image & video gen | Daniel Gilo +3 | Jun 18, 2026 | 21 |
| ImageWAM: Do World Action Models Really Need Video Generation, or Just Image Editing? | Robotics | Yuyang Zhang +9 | Jun 17, 2026 | 21 |
| Show the Signal, Hide the Noise: Spectral Forcing for Pixel-Space Diffusion | Image & video gen | Weichen Fan +3 | Jun 16, 2026 | 21 |
| Demystifying Hidden-State Recurrence: Switchable Latent Reasoning with On-Policy Reinforcement Learning | Reinforcement learning | Jiayu Yang +8 | Jun 11, 2026 | 21 |
| Breaking Entropy Bounds: Accelerating RL Training via MTP with Rejection Sampling | Agents | Yucheng Li +16 | Jun 10, 2026 | 21 |
| A Geometric Account of Activation Steering through Angle-Norm Decomposition | Safety & alignment | Georgii Aparin +1 | Jun 4, 2026 | 21 |
| Watch, Remember, Reason: Human-View Video Understanding with MLLMs | Speech & audio | Jiahao Meng +14 | Jun 5, 2026 | 21 |
| Rethinking Continual Experience Internalization for Self-Evolving LLM Agents | Agents | Jingwen Chen +9 | Jun 3, 2026 | 21 |
| Decentralized Instruction Tuning: Conflict-Aware Splitting and Weight Merging | Vision & multimodal | Minsik Choi +1 | Jun 1, 2026 | 21 |
| Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents | Agents | Minhua Lin +16 | May 28, 2026 | 21 |
| LongDS-Bench: On the Failure of Long-Horizon Agentic Data Analysis | Agents | Kewei Xu +6 | May 28, 2026 | 21 |
| Xetrieval: Mechanistically Explaining Dense Retrieval | Code | Zhixin Cai +9 | May 28, 2026 | 21 |
| KVpop -- Key-Value Cache Compression with Predictive Online Pruning | Code | Lukas Hauzenberger +7 | Jul 6, 2026 | 20 |
| ACID: Action Consistency via Inverse Dynamics for Planning with World Models | Robotics | Gawon Seo +2 | Jul 2, 2026 | 20 |
| Are We Measuring Strategy or Phrasing? The Gap Between Surface- and Approach-Level Diversity in LLM Math Reasoning | Reinforcement learning | Sangmook Lee +5 | Jun 29, 2026 | 20 |
| MultiHashFormer: Hash-based Generative Language Models | Data & benchmarks | Huiyin Xue +2 | Jun 26, 2026 | 20 |
| Semantic Browsing: Controllable Diversity for Image Generation | Image & video gen | Sara Dorfman +4 | Jun 22, 2026 | 20 |
| SkillHarness: Harnessing Safe Skills for Computer-Use Agents | Agents | Yurun Chen +3 | Jun 2, 2026 | 20 |
| JanusMesh: Fast and Zero-Shot 3D Visual Illusion Generation via Cross-Space Denoising | Image & video gen | Siang-Ling Zhang +3 | Jun 18, 2026 | 20 |
| The Reward Was in Your Data All Along: Correcting Flow Matching with Discriminator-Guided RL | Vision & multimodal | Nicolas Beltran-Velez +6 | Jun 17, 2026 | 20 |
| LLM Agents Can See Code Repositories | Vision & multimodal | Dongjian Ma +5 | Jun 12, 2026 | 20 |
| VideoMDM: Towards 3D Human Motion Generation From 2D Supervision | Image & video gen | Amir Mann +3 | Jun 11, 2026 | 20 |
| Workflow-GYM: Towards Long-Horizon Evaluation of Computer-use Agentic tasks in Real-World Professional Fields | Agents | Liya Zhu +39 | Jun 9, 2026 | 20 |
| UnpredictaBench: A Benchmark for Evaluating Distributional Randomness in LLMs | Data & benchmarks | Amirhossein Abaskohi +6 | Jun 4, 2026 | 20 |
| ZipSplat: Fewer Gaussians, Better Splats | Vision & multimodal | Alexander Veicht +3 | Jun 3, 2026 | 20 |
| PEEK: Picking Essential frames via Efficient Knowledge distillation | Vision & multimodal | Killian Steunou +4 | May 29, 2026 | 20 |
| Recovering Policy-Induced Errors: Benchmarking and Trajectory Synthesis for Robust GUI Agents | Agents | Tianpeng Bu +9 | May 28, 2026 | 20 |