HyWorldVLA混合世界模型以90.59 PDMS登顶NAVSIM，比亚迪AI团队首亮相
BYD AI Team Revealed for First Time: HyWorldVLA Hybrid World Model Achieves SOTA on NAVSIM v1 With 90.59 PDMS Marking BYD Entry Into Autonomous Driving Foundation Models
比亚迪悄悄拿出的HyWorldVLA，在NAVSIM上直接以90.59分拿下SOTA，这是他们首次秀自动驾驶基础模型肌肉。
比亚迪AI团队首次公开HyWorldVLA混合世界模型，采用像素-潜在混合世界建模与VLA架构，在NAVSIM v1基准上取得90.
59 PDMS，达到SOTA。该模型由比亚迪与哈尔滨工业大学机器人研究人员合作开发。
这一成绩标志着比亚迪正式进入自动驾驶基础模型赛道。
BYD AI Team Revealed for First Time: HyWorldVLA Hybrid World Model Achieves SOTA on NAVSIM v1 With 90.59 PDMS Marking BYD Entry Into Autonomous Driving Foundation Models
BYD AI team publishes HyWorldVLA with hybrid pixel-latent world modeling and VLA architecture, scoring 90.59 PDMS on NAVSIM benchmark, from team including HIT robotics researchers.
This report is categorized under AITOP's "Models" desk. Key entities and technical topics under continuous tracking include HyWorldVLA, 比亚迪, NAVSIM, 自动驾驶.
Primary source verification: published directly by pandaily with first-party authority.
