| Gemma 4 Technical Report | Speech & audio | Gemma Team +39 | Jul 2, 2026 | 60 |
| Multi-LCB: Extending LiveCodeBench to Multiple Programming Languages | Code | Maria Ivanova +7 | Jun 18, 2026 | 60 |
| Beyond Scalar Rewards by Internalizing Reasoning into Score Distributions | Image & video gen | Xin Jin +10 | Jun 8, 2026 | 60 |
| Why Far Looks Up: Probing Spatial Representation in Vision-Language Models | Vision & multimodal | Cheolhong Min +7 | May 28, 2026 | 60 |
| KVarN: Variance-Normalized KV-Cache Quantization Mitigates Error Accumulation in Reasoning Tasks | Code | Lorenz K. Muller +5 | Jun 2, 2026 | 59 |
| Representation Forcing for Bottleneck-Free Unified Multimodal Models | Image & video gen | Yuqing Wang +12 | May 29, 2026 | 59 |
| Self-Improving Language Models with Bidirectional Evolutionary Search | Robotics | Guowei Xu +6 | May 27, 2026 | 59 |
| Video-Oasis: Rethinking Evaluation of Video Understanding | Vision & multimodal | Geuntaek Lim +7 | Jul 2, 2026 | 58 |
| Formalizing Latent Thoughts: Four Axioms of Thought Representation in LLMs | Efficiency & systems | Fahd Seddik +1 | May 7, 2026 | 58 |
| GameCraft-Bench: Can Agents Build Playable Games End-to-End in a Real Game Engine? | Vision & multimodal | Tongxu Luo +24 | Jun 16, 2026 | 58 |
| minWM: A Full-Stack Open-Source Framework for Real-Time Interactive Video World Models | Image & video gen | Min Zhao +11 | May 28, 2026 | 58 |
| SwanVoice: Expressive Long-Form Zero-Shot Speech Synthesis for Both Monologue and Dialogue | Speech & audio | Ruiqi Li +5 | May 29, 2026 | 57 |
| Embodied.cpp: A Portable Inference Runtime of Embodied AI Models on Heterogeneous Robots | Robotics | Ling Xu +8 | Jul 2, 2026 | 56 |
| From Chatbot to Digital Colleague: The Paradigm Shift Toward Persistent Autonomous AI | Agents | Yongheng Zhang +19 | Jun 12, 2026 | 56 |
| Measuring Epistemic Resilience of LLMs Under Misleading Medical Context | Agents | Hongjian Zhou +21 | Jun 10, 2026 | 56 |
| K-BrowseComp: A Web Browsing Agent Benchmark Grounded in Korean Contexts | Agents | Nahyun Lee +14 | Jun 1, 2026 | 55 |
| OPID: On-Policy Skill Distillation for Agentic Reinforcement Learning | Agents | Shuo Yang +10 | Jun 25, 2026 | 54 |
| World Action Models: A Survey | Robotics | Qiuhong Shen +7 | Jun 18, 2026 | 54 |
| LabVLA: Grounding Vision-Language-Action Models in Scientific Laboratories | Robotics | Baochang Ren +17 | Jun 11, 2026 | 54 |
| Mellum2 Technical Report | Agents | Marko Kojic +8 | May 29, 2026 | 54 |
| YoCausal: How Far is Video Generation from World Model? A Causality Perspective | Image & video gen | You-Zhe Xie +5 | May 28, 2026 | 54 |
| ResearchStudio-Idea: An Evidence-Grounded Research-Ideation Skill Suite from ML Conference Outcomes | Data & benchmarks | Qihao Zhao +10 | Jul 5, 2026 | 53 |
| MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction | Robotics | Jianing Zhang +12 | Jun 17, 2026 | 53 |
| Where Do Deep-Research Agents Go Wrong? Span-Level Error Localization in Agent Trajectories | Agents | Jiaming Wang +10 | Jun 1, 2026 | 53 |
| Scenes as Objects, Not Primitives: Instance-Structured 3D Tokenization from Unposed Views | Robotics | Mijin Yoo +5 | Jun 28, 2026 | 52 |
| Escaping the Self-Confirmation Trap: An Execute-Distill-Verify Paradigm for Agentic Experience Learning | Agents | Shiding Zhu +8 | Jun 23, 2026 | 52 |
| ACE-Ego-0: Unifying Egocentric Human and Robotic Data for VLA Pretraining | Robotics | Hao Li +10 | Jun 15, 2026 | 52 |
| Retrospective Harness Optimization: Improving LLM Agents via Self-Preference over Trajectory Rollouts | Agents | Wenbo Pan +7 | Jun 4, 2026 | 52 |
| From Activation to Causality: Discovery of Causal Visual Representations in the Human Brain | Vision & multimodal | Yuval Golbari +7 | May 22, 2026 | 52 |
| Dual Latent Memory in Vision-Language-Action Models for Robotic Manipulation | Robotics | Hongyu Qu +8 | Jul 8, 2026 | 51 |
| Scaling Mixture-of-Experts Video Pretraining for Embodied Intelligence | Robotics | Shuailei Ma +26 | Jul 8, 2026 | 51 |
| TRL-Bench: Standardizing Cross-Paradigm Representation-Level Evaluation of Tabular Encoders | Code | Wei Pang +12 | Jun 8, 2026 | 51 |
| SoCRATES: Towards Reliable Automated Evaluation of Proactive LLM Mediation across Domains and Socio-cognitive Variations | Agents | Taewon Yun +5 | Jun 4, 2026 | 51 |
| Harness-1: Reinforcement Learning for Search Agents with State-Externalizing Harnesses | Agents | Pengcheng Jiang +7 | Jun 1, 2026 | 51 |
| DataComp-VLM: Improved Open Datasets for Vision-Language Models | Vision & multimodal | Matteo Farina +35 | Jun 26, 2026 | 50 |
| EvoPolicyGym: Evaluating Autonomous Policy Evolution in Interactive Environments | Agents | Zhilin Wang +15 | Jul 2, 2026 | 50 |
| PhysisForcing: Physics Reinforced World Simulator for Robotic Manipulation | Robotics | Peiwen Zhang +13 | Jun 26, 2026 | 50 |
| Playful Agentic Robot Learning | Robotics | Junyi Zhang +19 | Jun 17, 2026 | 50 |
| SearchSwarm: Towards Delegation Intelligence in Agentic LLMs for Long-Horizon Deep Research | Agents | Pu Ning +9 | Jun 8, 2026 | 50 |
| Qwen-Image-Agent: Bridging the Context Gap in Real-World Image Generation | Image & video gen | Zekai Zhang +20 | Jun 25, 2026 | 49 |
| ResearchMath-14K: Scaling Research-Level Mathematics via Agents | Agents | Guijin Son +5 | May 27, 2026 | 49 |
| Why Can't I Open My Drawer? Mitigating Object-Driven Shortcuts in Zero-Shot Compositional Action Recognition | Efficiency & systems | Geo Ahn +5 | Jul 2, 2026 | 48 |
| Morphing into Hybrid Attention Models | Efficiency & systems | Disen Lan +7 | Jun 29, 2026 | 48 |
| ReFreeKV: Towards Threshold-Free KV Cache Compression | Code | Xuanfan Ni +8 | Jun 26, 2026 | 48 |
| Qwen-Image-2.0-RL Technical Report | Image & video gen | Yixian Xu +27 | Jun 25, 2026 | 48 |
| Beyond the Current Observation: Evaluating Multimodal Large Language Models in Controllable Non-Markov Games | Vision & multimodal | Shengyuan Ding +6 | Jun 17, 2026 | 48 |
| CoVEBench: Can Video Editing Models Handle Complex Instructions? | Vision & multimodal | Jiangtao Wu +9 | Jun 7, 2026 | 48 |
| TUA-Bench: A Benchmark for General-Purpose Terminal-Use Agents | Agents | Shoufa Chen +9 | Jun 26, 2026 | 47 |
| The Verification Horizon: No Silver Bullet for Coding Agent Rewards | Agents | Binghai Wang +11 | Jun 24, 2026 | 47 |
| KaLM-Reranker-V1: Fast but Not Late Interaction for Compressed Document Reranking | Code | Xinping Zhao +9 | Jun 22, 2026 | 47 |
| ArcANE: Do Role-Playing Language Agents Stay in Character at the Right Time? | Agents | Woojung Song +5 | Jun 4, 2026 | 47 |
| ShutterMuse: Capture-Time Photography Guidance with MLLMs | Vision & multimodal | Jiayu Li +7 | Jun 24, 2026 | 46 |
| OpenThoughts-Agent: Data Recipes for Agentic Models | Agents | Negin Raoof +39 | Jun 23, 2026 | 46 |
| HarnessX: A Composable, Adaptive, and Evolvable Agent Harness Foundry | Agents | Tingyang Chen +13 | Jun 12, 2026 | 46 |
| GENEB: Why Genomic Models Are Hard to Compare | Safety & alignment | Daria Ledneva +2 | Jun 3, 2026 | 46 |
| DenoiseRL: Bootstrapping Reasoning Models to Recover from Noisy Prefixes | Reinforcement learning | Caijun Xu +3 | May 27, 2026 | 46 |
| Orchestra-o1: Omnimodal Agent Orchestration | Speech & audio | Fan Zhang +10 | Jun 10, 2026 | 45 |
| Vision as Unified Multimodal Generation | Image & video gen | Xiaoyang Han +16 | Jul 7, 2026 | 44 |
| MANCE: Manifold Aware Concept Erasure | Vision & multimodal | Matan Avitan +2 | Jul 4, 2026 | 44 |
| Beyond IID: How General Are Tabular Foundation Models, Really? | Data & benchmarks | Lennart Purucker +9 | Jun 29, 2026 | 44 |
| MMAE: A Massive Multitask Audio Editing Benchmark | Speech & audio | Ziyang Ma +37 | Jun 5, 2026 | 44 |
| TIDE: Proactive Multi-Problem Discovery via Template-Guided Iteration | Agents | Soyeong Jeong +3 | Jun 3, 2026 | 44 |
| GGT-100K: Generative Ground Truth for Generalizable Real-World Image Restoration | Vision & multimodal | Xiangtao Kong +4 | May 29, 2026 | 44 |
| SkillHone: A Harness for Continual Agent Skill Evolution Through Persistent Decision History | Agents | Zhiwei Li +1 | Jun 23, 2026 | 43 |
| Improved Large Language Diffusion Models | Image & video gen | Shen Nie +9 | Jun 24, 2026 | 43 |
| MobileForge: Annotation-Free Adaptation for Mobile GUI Agents with Hierarchical Feedback-Guided Policy Optimization | Vision & multimodal | Guangyi Liu +10 | Jun 18, 2026 | 43 |
| Vision Pretraining for Dense Spatial Perception | Robotics | Zelin Fu +8 | Jul 6, 2026 | 42 |
| Trimming the Long-Tail of Visual World Modeling Evaluation | Image & video gen | Bingxuan Li +8 | Jun 23, 2026 | 42 |
| MemGUI-Agent: An End-to-End Long-Horizon Mobile GUI Agent with Proactive Context Management | Vision & multimodal | Guangyi Liu +9 | Jun 18, 2026 | 42 |
| Beyond Uniform Token-Level Trust Region in LLM Reinforcement Learning | Reinforcement learning | Renjie Mao +9 | Jun 9, 2026 | 42 |
| Trust Region On-Policy Distillation | Agents | Xingrun Xing +4 | May 31, 2026 | 42 |
| How LoRA Remembers? A Parametric Memory Law for LLM Finetuning | Code | Ziwen Xu +6 | May 28, 2026 | 42 |
| PerceptionRubrics: Calibrating Multimodal Evaluation to Human Perception | Vision & multimodal | Yana Wei +16 | Jun 26, 2026 | 41 |
| Evolution Fine-Tuning: Learning to Discover Across 371 Optimization Tasks | Agents | Young-Jun Lee +7 | Jun 27, 2026 | 41 |
| Beyond Static Leaderboards: Predictive Validity for the Evaluation of LLM Agents | Vision & multimodal | Dhaval C. Patel +39 | Jun 18, 2026 | 41 |
| SCAIL-2: Unifying Controlled Character Animation with End-to-end In-Context Conditioning | Vision & multimodal | Wenhao Yan +3 | Jun 9, 2026 | 41 |
| SpatialWorld: Benchmarking Interactive Spatial Reasoning of Multimodal Agents in Real-World Tasks | Vision & multimodal | Hongcheng Gao +20 | Jun 8, 2026 | 41 |
| Function2Scene: 3D Indoor Scene Layout from Functional Specifications | Vision & multimodal | Ruiqi Wang +6 | May 29, 2026 | 41 |
| LongTraceRL: Learning Long-Context Reasoning from Search Agent Trajectories with Rubric Rewards | Agents | Nianyi Lin +3 | May 29, 2026 | 41 |
| GEM: Generative Supervision Helps Embodied Intelligence | Robotics | Ruowen Zhao +11 | May 27, 2026 | 41 |
| Geometry-Aware Representation Denoising for Robust Multi-view 3D Reconstruction | Image & video gen | Jin Hyeon Kim +10 | May 25, 2026 | 41 |
| Personalize-then-Store: Benchmarking and Learning Personalized Memory for Long-horizon Agents | Agents | Yeonjun In +4 | May 25, 2026 | 41 |
| ThriftAttention: Selective Mixed Precision for Long-Context FP4 Attention | Code | Joe Sharratt | May 21, 2026 | 41 |
| Multi-Block Diffusion Language Models | Image & video gen | Yijie Jin +10 | Jun 30, 2026 | 40 |
| Flow-DPPO: Divergence Proximal Policy Optimization for Flow Matching Models | Image & video gen | Bowen Ping +5 | Jun 9, 2026 | 40 |
| AdaPlanBench: Evaluating Adaptive Planning in Large Language Model Agents under World and User Constraints | Agents | Jiayu Liu +10 | Jun 4, 2026 | 40 |
| OrbitQuant: Data-Agnostic Quantization for Image and Video Diffusion Transformers | Image & video gen | Donghyun Lee +7 | Jul 2, 2026 | 39 |
| Translation as a Bridging Action: Transferring Manipulation Skills from Humans to Robots | Robotics | Sijin Chen +8 | Jun 26, 2026 | 39 |
| S-Agent: Spatial Tool-Use Elicits Reasoning for Spatial Intelligence | Vision & multimodal | Yalun Dai +12 | Jun 18, 2026 | 39 |
| LectūraAgents: A Multi-Agent Framework for Adaptive Personalized AI-Assisted Learning and Embodied Teaching | Robotics | Jaward Sesay +5 | Jun 15, 2026 | 39 |
| MemTrace: Tracing and Attributing Errors in Large Language Model Memory Systems | Code | Xinle Deng +17 | May 27, 2026 | 39 |
| The MiniMax-M2 Series: Mini Activations Unleashing Max Real-World Intelligence | Agents | MiniMax +206 | May 26, 2026 | 39 |
| $D^2$-Monitor: Dynamic Safety Monitoring for Diffusion LLMs via Hesitation-Aware Routing | Image & video gen | Aoxi Liu +7 | May 25, 2026 | 39 |
| ViQ: Text-Aligned Visual Quantized Representations at Any Resolution | Vision & multimodal | Xumin Yu +7 | Jun 25, 2026 | 38 |
| Beyond NL2Code: A Structured Survey of Multimodal Code Intelligence | Vision & multimodal | Xuanle Zhao +18 | Jun 16, 2026 | 38 |
| Unlimited OCR Works | Speech & audio | Youyang Yin +16 | Jun 22, 2026 | 38 |
| Kairos: A Native World Model Stack for Physical AI | Robotics | Kairos Team +23 | Jun 16, 2026 | 38 |
| MemDreamer: Decoupling Perception and Reasoning for Long Video Understanding via Hierarchical Graph Memory and Agentic Retrieval Mechanism | Vision & multimodal | Cong Chen +9 | Jun 5, 2026 | 38 |
| Humanoid-GPT: Scaling Data and Structure for Zero-Shot Motion Tracking | Data & benchmarks | Zekun Qi +12 | Jun 2, 2026 | 38 |
| Task-Focused Memorization for Multimodal Agents | Robotics | Tao Zou +4 | May 29, 2026 | 38 |