Chinese car news site Dongchedi and CCTV News ran a big test on driver assist systems (ADAS). They tried 36 different cars from popular brands, including start-ups, tech newcomers like Huawei, and old, well-known automakers. Each car went through 15 different real-life danger scenarios on highways and city streets. Tesla’s vision-only system came out on top. Many high-tech Chinese cars with lidar could not keep up.
Experts from public safety and Beihang University made sure rules matched China’s Level 2 ADAS standards.
Test Scenarios
The test covered six highway and nine city events. These included:
- A crashed car suddenly appears.
- Lane closed at night with a merging truck.
- Construction zone with a required lane shift.
- Lead car disappears (causing a false alarm).
- Tight merge with traffic entering the highway.
- Animal runs across the road (“reckless pig”).
Each car had one try per event. A pass meant staying in the lane or stopping in time without help from the driver and not braking too hard.
How Results Were Judged
Engineers gathered data from the cars. Independent people gave “pass” or “fail” for each scenario. The results were made public with a simple scorecard.
Results: Highway Test
Top Scorers
- Tesla Model 3 RWD (Vision only): 5/6
- Tesla Model X LR (Vision only): 5/6
- Xpeng G6 (with lidar): 3/6
- Wenjie M9 (with more lidar): 3/6
- Zhijie R7: 3/6
- BYD Z9GT EV: 3/6
Middle Group
- Aion RT (lidar): 2/6
- Platinum 3X: 2/6
- Avita 12 (heavy on sensors): 2/6
- Wenjie M7: 2/5
- Avita 07: 2/5
Low Scores
- Ideal L6: 1/6
- Xiaomi SU7: 1/5
- Wenjie M8: 1/5
- QinLDM: 1/5
- iCAR V23: 1/5
- Xiaomi SU7 Ultra: 1/4
- BYD Seagull: 1/4
- NIO ES6: 1/4
Cars That Failed All Highway Scenarios
- Zeekr 001 (2025 YOU Edition) – 0/6
- Baojun Enjoy PHEV – 0/6
- Lynk & Co – 0/6
- Han LEV – 0/5
- Zero runing C10 – 0/5
- PASSAT – 0/5
- GAC Honda P7 – 0/5
- Zeekr 7X (2025 100kWh) – 0/5
- Xpeng P7+ – 0/4
- Song Pro DM – 0/4
- Letao L60 – 0/4
- Star Era ET – 0/4
- Firefly – 0/4
Full scores and car details were provided in the public video and reports.

English Translation: 36 Car Models: ADAS Highway Accident Scenario Test – Full Results
Symbols: ✔ = Pass, X = Fail, – = Not tested.
Rank | Car Model | Highway Scenarios Passed / Total | Sudden Accident Vehicle | Lane Closed | Road Construction | Lead Car Disappears | Animal Cross | Ramp Merge |
---|---|---|---|---|---|---|---|---|
1 | Tesla Model 3 (2023 RWD Refreshed) | 5 / 6 | ✔ | X | ✔ | ✔ | X | ✔ |
2 | Tesla Model X (2023 LR Refreshed) | 5 / 6 | ✔ | X | ✔ | ✔ | ✔ | X |
3 | Xpeng G6 (2024 Max) | 3 / 6 | ✔ | X | ✔ | X | X | ✔ |
4 | Wenjie M9 (AITO, 2024) | 3 / 6 | X | ✔ | ✔ | X | X | ✔ |
5 | Zhijie R7 (Avatar R7) | 3 / 6 | ✔ | X | ✔ | X | X | ✔ |
6 | BYD Denza Z9GT EV (2024) | 3 / 6 | ✔ | ✔ | X | X | X | ✔ |
7 | GAC Aion RT (2025) | 2 / 6 | ✔ | X | ✔ | X | X | X |
8 | Chery Fulwin 3X (2024) | 2 / 6 | ✔ | X | ✔ | X | X | X |
9 | Changan Avatr 12 (2024) | 2 / 6 | ✔ | X | ✔ | X | X | X |
10 | Wenjie M7 (AITO, 2023 Ultra) | 2 / 5 | ✔ | X | ✔ | X | X | – |
11 | Avita 07 (2025 Ultra) | 2 / 5 | ✔ | X | ✔ | X | X | – |
12 | Li Auto L6 (2024 Ultra Max) | 1 / 6 | ✔ | X | X | X | X | X |
13 | Xiaomi SU7 (2024 Max) | 1 / 5 | ✔ | X | X | X | X | – |
14 | Wenjie M8 (AITO) | 1 / 5 | ✔ | X | X | X | X | – |
15 | BYD Qin PLUS DM-i | 1 / 5 | ✔ | X | X | X | X | – |
16 | iCAR V23 (2025 SUV) | 1 / 5 | ✔ | X | X | X | X | – |
17 | Zeekr X (2024) | 1 / 4 | ✔ | X | X | X | – | – |
18 | BYD Seagull (2024 305KM) | 1 / 4 | ✔ | X | X | X | – | – |
19 | NIO ES6 (2024 100kWh) | 1 / 4 | ✔ | X | X | X | – | – |
20 | Xiaomi SU7 Ultra (2025) | 1 / 4 | ✔ | X | X | X | – | – |
21 | Zeekr 001 (2025 YOU Edition) | 0 / 6 | X | X | X | X | X | X |
22 | Baojun Enjoy PHEV (2025) | 0 / 6 | X | X | X | X | X | X |
23 | Lynk & Co 08 | 0 / 6 | X | X | X | X | X | X |
24 | Han LEV (BYD) | 0 / 5 | X | X | X | X | X | – |
25 | GAC Honda P7 (2025) | 0 / 5 | X | X | X | X | X | – |
26 | Passat (Volkswagen, 2025) | 0 / 5 | X | X | X | X | X | – |
27 | Xpeng P7+ | 0 / 4 | X | X | X | X | – | – |
28 | BYD Song Pro DM-i (2025) | 0 / 4 | X | X | X | X | – | – |
29 | Jetour X90 (2025 Ultra) | 0 / 4 | X | X | X | X | – | – |
30 | Firefly (2025) | 0 / 4 | X | X | X | X | – | – |
… | … | … | … | … | … | … | … | … |
Why Tesla Scored Higher
Vision-Only Outperformed Lidar
Tesla uses only cameras for its ADAS. These models scored higher than most cars with lots of sensors. Reasons:
- Tesla trains its AI on huge amounts of driving data, which helps the car spot hard-to-see dangers.
- The vision system responds quickly, letting it react to fast changes like animals or sudden obstacles.
- Tesla’s system links what it sees with how the car moves, making errors less likely.
Chinese cars with lidar often made mistakes. For example, some mistook animals for road trash because the sensors got confused or missed them.
Lidar Limitations
Some cars using lidar lost sight of hazards, especially if objects were small or partially blocked. This caused them to miss events, like when a lead car vanished or when the test pig crossed the road.
Data Gaps for Chinese Start-Ups
Many brands built their car’s computer brains using mostly city driving data. This left a hole for scenarios like animals on highways, which led to many failures. Tesla’s global data collection gave it an advantage, even in China.
More Gear Didn’t Always Help
Cheaper cars sometimes did better than pricier ones. Some models with lots of expensive hardware failed most scenarios because their software wasn’t advanced enough.
Urban Test Snapshot
Tesla Model X also did well in city scenarios, passing 8 out of 9. Huawei-backed brands were just behind. No car had a perfect score. On average, success across all tests was about 36%.
Traffic safety officials told drivers that all these systems are driver-assist only. Drivers must pay attention and keep their hands on the wheel.
Automaker Responses
- Tesla pointed out their cars used no local Chinese training data.
- Huawei and AITO, who use layered sensor systems, did not comment much but said they would release upgraded features later in the year.
- Some Chinese brands promised future software updates and said they welcome further testing.
Some people said one try per scenario isn’t enough for perfect accuracy. Others argued these open tests are better than marketing claims without real data.
Why Tesla Missed the Animal Test
The test vehicle saw the animal run across but could not stop in time at highway speed. The driver had to step in.
Lidar Blind Spots
Other cars lost track of the animal because it was hidden or in a shadow. Their sensors flagged something, but the car’s computer filtered out the warning.
How ADAS Systems Compare
Tesla’s simple hardware and strong software beat most rivals using more sensors. Some Chinese brands may switch to simpler systems or mix vision AI with lidar to catch up.
However, Driver assist is not self-driving. Stay alert and ready to take over at any time. Software updates can improve these systems. If you drive in places with animal crossings, check that your car can handle it; most still cannot.
Dongchedi said they will run another test in late 2025. This next one will focus on night and rain. Tesla may update its hardware before then, and Huawei plans changes to its ADAS, hoping to improve.
The test showed that real driving data and fast, smart software are more important than just adding more sensors. This gap in software ability, not hardware, is now the main challenge for automakers.
Dongchedi led the project and shared a 1.5-hour video. You can watch it below:
You may also like to read:
- Tesla Moves Closer to Affordable Electric Cars with June Pilot Models »
- Tesla Releases Q2 2025 Safety Report: Autopilot Now 9.5 Times Safer Than Manual Driving »
- Model 3+ Will Offer Over 800 km Range, Tesla Prepares Launch »