Qianli Smart Driving Moves Into Volume Ramp With RMB100,000-500,000 Coverage
Qianli Smart Driving enters volume phase as two Galaxy models extend it to the RMB100,000-200,000 band, Leiphone reports.
According to the report, the two models are not merely additional production cars. They mark the company's move from high-end capability validation to mainstream-scale application. In the first half of 2026, Qianli's systems covered Zeekr, Lynk & Co and smart models; the Galaxy launches extended that coverage into the RMB100,000-200,000 price band. The report describes Qianli's path as high-end start, mainstream reuse and scale ramp, with earlier production models including Zeekr 9X, Zeekr 8X, Lynk & Co 07GT and smart #1.
The report places Qianli's push against two industry windows. AI large models are moving driving systems from rule-driven software toward model-driven systems, raising the ceiling for perception, reasoning and decision-making in complex scenes. At the same time, competition is shifting from features and performance toward mass production and scale, with entry into mainstream price bands and wider model coverage becoming central. Qianli's answer rests on three supports: market, model and data.
On the market side, Qianli entered later but has covered price bands from high-end to mainstream through models for Zeekr, Lynk & Co and smart. Geely Galaxy sold nearly 1.24 million vehicles in 2025 and more than 800,000 in the first three quarters of 2026, giving Qianli a base for cross-model and cross-platform deployment, according to the report. The company follows an AI-plus-car route, using AI model capabilities as its core, validating high-end production with models such as Zeekr 9X and connecting chip, computing and lidar partners through an open platform. The report cites strategic automaker customers, AI-native origins and chain collaboration as its position in the industry.
On the model side, Qianli Technology has formed a strategic cooperation with StepFun to build a native driving base model under a brain-body collaboration approach. The model combines general corpus and driving data during pretraining, aiming at recognition, reasoning and decision-making in real road environments. Qianli uses a unified L4 architecture and releases some L4 technologies into production models, so one technical base can support products at different positions. In September 2026, Qianli Smart Driving, together with the University of Electronic Science and Technology of China, Southeast University, Beijing University of Posts and Telecommunications and Tianjin University, released and open-sourced the World-Action model WA-JEPA, listed as arXiv:2608.20974.
On the data side, the report says the industry is shifting from comparing functions to comparing iteration. Production scale becomes material for model evolution: real mileage and long-tail scenes are collected through collection vehicles, test vehicles and production vehicles, trained and validated on supercomputing clusters, and fed back into models in a day-level data loop. As mainstream models ramp, more data is expected to drive model and experience upgrades, wider vehicle installations and more user use, forming a cycle in which scale produces data, data trains models, models improve experience and experience drives growth.
The Galaxy TT and Galaxy Warship 700 launches are an important node but not the end, according to the report. Qianli is moving from single-model production toward a brand matrix and multi-price-band scale. The larger change is full integration into the Galaxy brand system: Zeekr, Lynk & Co and smart have validated capability and reputation, while Galaxy connects that capability to a brand system with annual sales above one million vehicles. That means installations within Galaxy can expand with a sustained volume base rather than through model-by-model negotiations. Citing Yiche Bang data for the first half of 2026, the report says Qianli's city NOA share in the above-RMB400,000 high-end market reached 40.8%, and its overall L2+ share reached 39.2%, both ranking first. The next observation window is city NOA installations in the RMB100,000-200,000 mainstream band as Galaxy models ramp. The report also describes Qianli Technology as more than an algorithm supplier: it connects chip ecosystems, computing, vehicles and user scenarios around AI models, covers intelligent driving and intelligent cockpits, and is building a platform that spans high-end and mainstream price bands and evolves from assisted driving toward L4.