What happened
The China Automotive Technology and Research Center (CATARC) rates assisted driving in C-ICAP, its own program. The revised 2024 C-ICAP rules state that once the models are chosen, CATARC's Management Center "will independently acquire them based on random principles." CATARC's own yearly budget pays for buying the cars and for the tests. On 4 June 2026 CATARC published its results for a long-range rear-wheel-drive Xiaomi YU7 with Hyper OS 1.12.0. Its rating report dates the tests from March to May 2026. The car scored 97.9% for basic driving assistance, with 100% for car following and combined control, 99.2% for emergency avoidance and 92.0% for driver interaction. The page also lists 97.8% for highway navigation assistance. The revised rules make that an audit item, judged from a qualified third-party lab's test report that the carmaker provides. CATARC also scores some simulated scenarios from reports that the carmaker supplies. The result covers one car and one software version, and no outcome in everyday driving is reported.
Sources
-
Assessment report, by authors with no tie to the system.
组合控制能力(得分率:100%) 紧急避险(得分率:99.2%) 驾驶员交互(得分率:92.0%) 高速领航辅助(得分率:97.8%)
Details
AI system
- Reported configuration
-
Hyper OS 1.12.0 (software version tested)
- Workflow
-
CATARC acquired the car itself and chose the configuration with the highest level of intelligence after asking the carmaker for technical information. The car ran Hyper OS 1.12.0 with an NVIDIA Thor chip, 11 cameras, 1 millimeter-wave radar, 12 ultrasonic radars and 1 lidar. CATARC tested the car from March to May 2026 under the revised 2024 rules.
Follow-up
- Conditions for replication
-
The driver stays in charge, since the page warns that assisted driving carries risk and that the driver bears first responsibility. The result applies only to the tested sample, software version and test environment.
Entry details
- Sector
-
Road vehicle crash avoidance / assisted driving.
- Purpose
-
Xiaomi built the assisted driving in Hyper OS to help the driver control speed, following distance and steering and to follow a navigation route on highways.
30 more questions not yet checked
model, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). Hyper OS 1.12.0 assisted driving scores 97.9% in basic tests on a YU7 (AIBD-1020). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-1020.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)