Workflow
行为强化学习
icon
Search documents
理想汽车推送OTA 8.2升级,AD Max平台VLA强化版上线
Bei Ke Cai Jing· 2026-01-21 10:17
Group 1 - The core focus of the news is the official rollout of OTA 8.2 upgrade by Li Auto on January 21, which includes the launch of the enhanced VLA version of the AD Max platform [1] - The upgrade emphasizes model reinforcement training, significantly reducing mechanical sensations during driving, leading to a smoother experience [1] - Li Auto's Senior Vice President of Autonomous Driving R&D, Lang Xianpeng, stated that extensive testing has shown the performance of the upgrade to reach industry-leading levels [1] Group 2 - The VLA driver model now incorporates behavior reinforcement learning capabilities, enhancing deep learning with millisecond-level data on steering and throttle actions [4] - The improvements result in more natural braking and turning, providing a more human-like driving experience, which contributes to a smoother and more comfortable ride [4] - The upgrade is applicable to all Li Auto AD Max models, including the Li MEGA, Li i series, and Li L series [6]
理想汽车OTA 8.2正式推送,引入VLA司机大模型行为强化学习
Feng Huang Wang· 2026-01-21 08:20
Core Viewpoint - Li Auto has launched the OTA 8.2 software update, introducing the "VLA driver model behavior reinforcement learning" capability to enhance vehicle dynamics and driving experience [1] Group 1: Software Update Features - The OTA 8.2 update includes 40 new and optimized features, such as the HUD narrow road imaging system that displays wheel hub perspective during narrow road navigation [1] - The update also introduces a pet mode, an art frame feature, and charging preheating for extended-range models [1] Group 2: Company Philosophy - Li Auto emphasizes its philosophy of "a mobile home, always new" through this software upgrade, aiming to continuously improve user experience [1]