Home: Motoring > QCraft to Deliver 21 Smart Driving Models by 2026, Spanning $11K–$57K Price Range and Launching World Model Upgrade

QCraft to Deliver 21 Smart Driving Models by 2026, Spanning $11K–$57K Price Range and Launching World Model Upgrade

From:Internet Info Agency 2026-04-07 11:12:10

QCraft announced that, as of 2026, 21 new vehicle models equipped with its intelligent driver-assistance systems have entered the delivery phase. Its automotive partners span private Chinese brands, central state-owned enterprises, leading automakers in South China, and industrial conglomerates venturing into the automotive sector. These models cover a broad price range from RMB 80,000 to over RMB 400,000, achieving mass production across multiple price segments. The company’s “QCraft Breeze 2.0” three-tier product matrix has been fully deployed: - The Breeze AIR solution features approximately 80 TOPS of computing power and a 7V sensor configuration, supporting highway NOA (Navigate on Autopilot). - The Breeze PRO solution offers 100–200 TOPS of computing power with either an 11V or 11V1L sensor setup, enabling urban NOA. - The Breeze MAX solution delivers over 500 TOPS of computing power with an 11V1L configuration, deeply integrating Vision-Language-Action (VLA) models and world models to support advanced urban NOA. Its single-chip urban NOA solution, powered by the Journey 6M chip, effectively handles diverse scenarios including urban arterials, rural roads, complex intersections, roundabouts, and U-turns. To date, the “QCraft Breeze” series has accumulated over 2.5 billion kilometers of user-assisted driving mileage, recorded nearly 100 million smart parking usages, and achieved an AEB (Automatic Emergency Braking) false-trigger rate of less than once per 400,000 kilometers—preventing an estimated 146,000 potential accidents annually. In January 2026, QCraft unveiled its unified “VLA + World Model” architecture, designed to enhance the system’s understanding of real-world complexity and advance capabilities in behavioral generalization, reasoning, and comprehension of physical laws and social common sense—laying the groundwork for Level 4 autonomous driving. At the 2026 Beijing Auto Show, the company will unveil its latest technical upgrade, integrating world models with reinforcement learning. QCraft currently possesses mass-production capacity at the million-unit scale and has successfully implemented its single-chip urban NOA solution in volume production.

Editor:NewsAssistant