Li Auto releases new autonomous driving model MindVLA-o1

At NVIDIA GTC 2026, Li Auto unveiled its next-generation autonomous driving foundation model, MindVLA-o1. This model unifies vision, language, and action within a single framework, marking a shift in autonomous driving technology towards a general-purpose intelligent agent. Li Auto's assisted driving technology architecture has undergone multiple iterations since 2021. The end-to-end + VLM dual-system architecture launched in 2024 achieved unified understanding capabilities across scenarios and tasks. In 2025, Li Auto launched the VLA driver master model, which achieved a monthly usage rate of 80% by the end of the year.