1
“VLA和世界模型都不是终局，会有物理世界独有的模型” | 蚂蚁灵波沈宇军@AIGC2026
蚂蚁灵波沈宇军提出AI 2.0下半场应从内容生成转向动作生成，认为VLA和世界模型都不是终局，物理世界需要独特模型。
“要做机器人时代的安卓系统”
2
VLANeXt: Recipes for Building Strong VLA Models
顶级会议ICML 2026收录，揭秘构建强视觉-语言-动作（VLA）模型的实用配方与技巧。
arXiv:2602.18532v2 Announce Type: replace Abstract: Following the rise of large foundation models, Vision-Language-Action models (VLAs) emerged, lever…
3
FASTER: Rethinking Real-Time Flow VLAs
聚焦实时流VLA架构创新，重新思考并加速Flow VLA推理效率，适合AI研究者。
arXiv:2603.19199v3 Announce Type: replace-cross Abstract: Real-time execution is crucial for deploying Vision-Language-Action (VLA) models in the phys…
4
Rethinking Muon Beyond Pretraining: Spectral Failures and High-Pass Remedies for VLA and RLVR
Muon优化器在视觉语言对齐与强化学习微调中暴露频谱失效问题，作者提出高通滤波器补救方案，刷新大模型训练认知。
arXiv:2605.19282v1 Announce Type: new Abstract: Muon is a matrix-aware optimizer that leverages Newton-Schulz (NS) iterations to enforce spectral grad…
5
Overcoming Dynamics-Blindness: Training-Free Pace-and-Path Correction for VLA Models
VLA模型存在动力学盲点？这篇论文提出无需训练的步速与路径校正方法，突破单帧观测局限。
arXiv:2605.11459v2 Announce Type: replace-cross Abstract: Vision-Language-Action (VLA) models achieve remarkable flexibility and generalization beyond…
6
UAM: A Dual-Stream Perspective on Forgetting in VLA Training
揭示VLA训练中VLM多模态能力系统退化的“具身税”现象，提出双流新视角UAM
arXiv:2605.15735v1 Announce Type: cross Abstract: Vision--language--action (VLA) models are typically built by fine-tuning a pretrained vision--langua…
7
Learn Where Outcomes Diverge: Efficient VLA RL via Probabilistic Chunk Masking
提出概率块掩码机制，直击VLA强化学习后训练计算瓶颈，显著提升效率。
arXiv:2605.16154v1 Announce Type: new Abstract: Reinforcement learning (RL) allows vision-language-action (VLA) policies to generalize beyond their tr…
8
何小鹏：小鹏 Robotaxi 将在广州开启示范运营
小鹏首台量产Robotaxi下线并将在广州示范运营，搭载自研图灵AI芯片与VLA大模型，实现L4级自动驾驶。
IT之家 5 月 18 日消息，今天，小鹏首台量产的 Robotaxi 在广州工厂下线了。小鹏集团董事长、CEO 何小鹏今日宣布， 未来几个月，小鹏 Robotaxi 会在广州开启示范运营 。 IT之家从官方介绍获悉，作为中国首款全栈自研、前装量产的 Robotaxi， 基于小鹏旗舰车型 GX 打造…