At the 2026 XPENG Global New Product Launch Event, He Xiaopeng, Chairman of XPENG Motors, declared that the Ultra variants of the 2026 XPENG P7+, G6, G7, and G9 models would come equipped with XPENG's second-generation Vision-Language-Action (VLA) model. This milestone positions XPENG as the industry's trailblazer, introducing the first large-scale model tailored for the physical world, boasting initial Level 4 (L4) autonomous driving capabilities. The model seamlessly integrates action generation with advanced physical world understanding and reasoning abilities. It supports self-evolving learning mechanisms, enabling it to operate a diverse array of vehicles across various domains. Leveraging a vast dataset of nearly 100 million real-world driving video clips, the model can be directly trained to enhance its performance. It excels in reasoning through future decision-making scenarios and generating long-tail scenarios for adversarial training, thereby significantly boosting its capacity to tackle complex challenges. Moreover, XPENG Motors is set to introduce the 'Xiaolu NGP' feature and has plans to roll out 'Navigation-Free Automatic Assisted Driving' with Super LCC+ human-machine co-driving capabilities.

Each new model boasts its distinct positioning in the market.