Nio said on Friday that the rollout of the first version of its “Nio World Model” (NWM) assisted driving system will begin this weekend for vehicles built on its latest Cedar operating system.
The EV maker initially released its smart driving system to older vehicles in late May and scheduled the rollout to its newer models for late June.
Starting this weekend, the NWM will be pushed over the air to the recently refreshed ES6, EC6, ET5, and ET5 Touring models.
The EV maker said the NWM system is designed to reduce accidents and driver workload across four major driving scenarios: active safety, highway pilot, urban navigation, and intelligent parking.
The Model includes forward collision warning, automatic emergency braking, lane-keeping functions, pedestrian protection, adaptive cruise control, and automated parking.
The update follows the initial rollout of the NWM to vehicles based on the older Banyan platform in late May.
That release marked the first large-scale deployment of the Nio World Model, introduced a year ago alongside the tape-out of the NX9031 chip, with Nio claiming to be the first Chinese automaker a smart driving model.
In a demonstration shared in mid-May, Harry Wong, Nio’s Head of Smart Driving Product and Experience, posted footage of a nearly 40-kilometre autonomous drive completed without any driver intervention.
NWM is a multimodal autoregressive model that integrates visual, linguistic, spatial, and temporal data to enable real-time reasoning and decision-making, according to the carmaker.
The first version introduces a “mapless” autonomous navigation function for complex environments such as underground parking lots.
Unlike conventional systems that rely on preloaded maps or fixed interface commands, NWM allows drivers to use natural language to instruct the vehicle—such as “take me to the exit” or “go to Building 18.”
The system interprets signage and surroundings using onboard cameras and sensors to generate an appropriate driving path.
The company described NWM as a foundational step in enabling point-to-point autonomous navigation without pre-defined routes.
“Parking Lot Autonomous Navigation” is the first application of the model, combining spatial perception, environmental understanding, and multi-turn reasoning within a map-free setting.
