en.Wedoany.com Reported - On June 13, 2026, the 28th Chongqing International Auto Show opened at the Chongqing International Expo Center. QCraft showcased its core Physical AI technology based on "World Model + Reinforcement Learning," along with its urban NOA solution and unmanned logistics vehicle products developed under this technical framework.
This year's auto show established the first Future Automotive AI Technology Exhibition, focusing on the integration of AI with the automotive industry, attracting multiple Chinese intelligent driving companies. Under the theme "Physical AI, QCraft Has Arrived," QCraft showcased its technical architecture, including a cloud-based world model and an on-vehicle world behavior model, as well as the Chengfeng MAX urban NOA solution based on over 500 TOPS of computing power and the Q5 unmanned logistics vehicle.
QCraft has built a core Physical AI architecture comprising a "cloud-based world model + on-vehicle world behavior model." The cloud-based world model is equipped with a world simulation engine and safe reinforcement learning, enabling high-controllable video generation, zero-shot scene generation, and low-cost closed-loop simulation capabilities. It can generate traffic flows, obstacle edge cases, high-risk rare scenarios, and adverse weather conditions through motion simulation, facilitating low-cost, high-fidelity long-tail scenario training. The on-vehicle world behavior model adopts the same technical approach, using the world model as its backbone to achieve end-to-end modeling from perception to decision-making, enhancing reasoning, decision-making, and generalization capabilities, covering passenger cars, Robotaxis, Robovans, and other vehicle types.
The Chengfeng MAX solution, based on Physical AI technology, employs a unified architecture model and supports urban NOA functions with over 500 TOPS of onboard computing power. This solution can predict the intentions of traffic participants, handle complex right-of-way negotiations, and identify pedestrian intentions in scenarios such as narrow village roads and complex intersections, making dynamic decisions that reflect defensive driving characteristics, thereby improving traffic efficiency and safety.
In the field of L4 autonomous driving, QCraft is advancing the dual-track development of Robotaxi and Robovan. The Robotaxi solution builds an AI brain based on the world model, optimizing decision-making through cloud-based simulation and reinforcement learning, establishing a "perception-prediction-action" general intelligence, with the goal of providing a mass-producible L4 mobility solution. The L4 unmanned logistics solution is adaptable to loading spaces ranging from 3 to 12 cubic meters, covering scenarios such as express delivery, industrial logistics, cold chain, instant retail delivery, and security patrols. It has already undergone large-scale testing and operations in multiple cities including Ningbo, Wuhu, and Jinhua. Additionally, QCraft has been conducting daily road tests for intelligent assisted driving in Munich and Paris.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
