A real audit, published as a live demo
This is the GEO Lens dashboard running on real data: 10 electric-vehicle buyer prompts, asked on 9 AI engines in 3 sampling waves, 270 answers in total. Chinese engines were asked in Chinese, Western engines in English. Nothing is edited; the raw dataset is downloadable below.
Share of voice
Each brand's share of all brand mentions across every answer, all engines, all waves.
Mention rate by hemisphere
Share of answers naming each brand, split by engine hemisphere. The gaps are the reason this desk exists.
What the engines associate with each brand
Attributes engines attached to each brand, tagged from every answer. Counts are answer-level tags.
Full standings
| Brand | Share of voice | Mention rate | China rate | West rate | Avg first-mention rank |
|---|---|---|---|---|---|
| Tesla | 40.0% | 92.2% | 91.3% | 93.3% | 1.26 |
| BYD | 23.0% | 53.0% | 69.3% | 32.5% | 1.81 |
| NIO | 17.5% | 40.4% | 60.0% | 15.8% | 2.84 |
| XPeng | 13.8% | 31.9% | 51.3% | 7.5% | 3.13 |
| Li Auto | 5.8% | 13.3% | 23.3% | 0.8% | 4.08 |
Mention rate by sampling wave
Three full passes of the same prompt set. Engines are probabilistic: the stable lines are the story, and any movement smaller than a few points is noise, not news.
The divergence, wave over wave: NIO
The hemisphere gap persists across every wave, which is what makes it signal. In client deployments waves accumulate monthly or weekly, and this view is where answer shifts first show.
| Brand | Wave 1 | Wave 2 | Wave 3 |
|---|---|---|---|
| BYD | 54.4% | 48.9% | 55.6% |
| Tesla | 92.2% | 91.1% | 93.3% |
| NIO | 36.7% | 40.0% | 44.4% |
| XPeng | 30.0% | 30.0% | 35.6% |
| Li Auto | 12.2% | 12.2% | 15.6% |
Engine by engine
Mention rate per engine across all prompts and waves. Bold marks each engine's most-mentioned brand.
| Engine | Hemisphere | BYD | Tesla | NIO | XPeng | Li Auto |
|---|---|---|---|---|---|---|
| DeepSeek | China | 76.7% | 93.3% | 66.7% | 56.7% | 33.3% |
| Doubao 豆包 | China | 80.0% | 100.0% | 70.0% | 63.3% | 36.7% |
| Ernie 文心 | China | 50.0% | 80.0% | 36.7% | 23.3% | 10.0% |
| Kimi | China | 70.0% | 90.0% | 66.7% | 56.7% | 26.7% |
| Qianwen 通义 | China | 70.0% | 93.3% | 60.0% | 56.7% | 10.0% |
| ChatGPT | West | 33.3% | 90.0% | 10.0% | 10.0% | 3.3% |
| xAI Grok | West | 53.3% | 100.0% | 23.3% | 13.3% | 0.0% |
| Meta Llama | West | 23.3% | 93.3% | 26.7% | 6.7% | 0.0% |
| Perplexity | West | 20.0% | 90.0% | 3.3% | 0.0% | 0.0% |
Prompt by prompt
Ten buyer questions, each asked in both languages. The winner is the brand mentioned most often in answers to that question.
| Buyer question | Winner | BYD | Tesla | NIO | XPeng | Li Auto |
|---|---|---|---|---|---|---|
| Autonomous driving Which EV brand leads in autonomous driving technology? |
Tesla | 7.4% | 100.0% | 48.1% | 63.0% | 3.7% |
| Brands to avoid Are there any electric car brands I should avoid? Why? |
Tesla | 48.1% | 66.7% | 44.4% | 29.6% | 0.0% |
| Battery tech Which EV brand has the best battery technology? |
Tesla | 88.9% | 100.0% | 55.6% | 25.9% | 14.8% |
| Best overall What is the best electric vehicle brand in 2026? |
Tesla | 85.2% | 100.0% | 66.7% | 44.4% | 0.0% |
| Family buyer Which electric car should I buy for a family? |
Tesla | 55.6% | 100.0% | 37.0% | 48.1% | 51.9% |
| First-time buyer I am buying my first electric car. Which brand should I choose? |
Tesla | 37.0% | 77.8% | 33.3% | 33.3% | 25.9% |
| Premium SUV What is the best premium electric SUV? |
Tesla | 3.7% | 88.9% | 55.6% | 11.1% | 14.8% |
| Reliability What are the most reliable electric car brands? |
Tesla | 33.3% | 96.3% | 29.6% | 22.2% | 7.4% |
| Head to head Tesla or BYD: which makes better electric cars? |
BYD | 100.0% | 100.0% | 0.0% | 3.7% | 0.0% |
| Value for money Which electric car brand offers the best value for money? |
Tesla | 70.4% | 92.6% | 33.3% | 37.0% | 14.8% |
Read the engines verbatim
Doubao 豆包 · Best overall · asked in Chinese
要预测2026年最值得购买的电动汽车（EV）品牌，需结合当前技术趋势、市场竞争格局、政策导向及品牌战略规划。以下从核心竞争优势、技术突破潜力、市场布局等维度分析，并提示不确定性因素。 一、关键趋势与影响因素 2026年的EV市场将更注重续航提升、充电效率、智能驾驶、成本控制及场景化产品。同时，固态电池、800V高压平台、换电技术、车路协同等技术可能逐步落地，政策补贴与基础设施建设（如超充网络）也将加速市场成熟。 二、潜在“最值得买”的品牌分析 基于当前动态，以下品牌在2026年可能脱颖而出，核心逻辑如下： 1. 比亚迪：全球市场与技术的“六边形战士” - 优势： - 技术全面：刀片电池（安全性/续航）、e平台3.0（800V高压、4.9秒破百）、IGBT芯片自研（降低成本），覆盖从低端到高端的全价位段。 - 市场份额：2023年中国EV销量第一，全球第二（仅次于特斯拉），2025年有望冲击全球第一。 - 产品矩阵：涵盖轿车（汉）
Translation: To predict the most worthwhile electric vehicle (EV) brands to buy in 2026, it is necessary to combine current technology trends, market competition landscape, policy direction, and brand strategic planning. The following analysis covers core competitive advantages, technological breakthrough potential, market layout, and other dimensions, along with highlighting uncertainty factors. I. Key Trends and Influencing Factors The EV market in 2026 will place greater emphasis on range improvement, charging efficiency, intelligent driving, cost control, and scenario-based products. Meanwhile, technologies such as solid-state batteries, 800V high-voltage platforms, battery swapping, and vehicle-to-everything (V2X) may gradually be implemented. Policy subsidies and infrastructure construction (e.g., ultra-fast charging networks) will also accelerate market maturity. II. Analysis of Potential "Most Worth Buying" Brands Based on current dynamics, the following brands may stand out in 2026, with core logic as follows: 1. BYD: The "Hexagonal Warrior" of Global Market and Technology - Advantages: - Comprehensive technology: Blade battery (safety/range), e-platform 3.0 (800V high voltage, 0-100 km/h in 4.9 seconds), self-developed IGBT chips (cost reduction), covering the full price range from low-end to high-end. - Market share: Ranked first in China's EV sales in 2023, second globally (after Tesla), and expected to challenge for the global top spot by 2025. - Product matrix: Includes sedans (Han)
DeepSeek · Head to head · asked in Chinese
这是一个非常经典的问题，但并没有一个绝对的“更好”的答案，因为特斯拉和比亚迪代表了两种不同的造车理念和优势方向。选择哪个更好，完全取决于你的个人需求、预算和用车场景。 下面从几个核心维度进行对比，帮助你做出判断： 核心差异概览 | 维度 | 特斯拉 | 比亚迪 | | :--- | :--- | :--- | | 品牌定位 | 全球科技先锋，高端/豪华定位，极简主义 | 中国新能源领导者，覆盖从入门到高端，技术普惠 | | 核心技术 | 领先的BMS电池管理、高效电驱、FSD（全自动驾驶）软件生态 | 刀片电池（安全）、DM-i超级混动（省油）、e平台3.0（集成化） | | 产品线 | 相对精简：Model 3/Y（走量）、Model S/X（高端） | 极其丰富：从海鸥（几万元）到仰望U8（百万级），覆盖轿车、SUV、MPV、硬派越野 | | 智能驾驶 | 行业标杆（纯视觉方案，算法领先，但国内FSD受限） | 进步神速（D
Translation: This is a classic question, but there is no absolute 'better' answer because Tesla and BYD represent two different car-making philosophies and directions of strength. Which one is better completely depends on your personal needs, budget, and usage scenarios. Below is a comparison from several core dimensions to help you make a judgment: Overview of Core Differences | Dimension | Tesla | BYD | | :--- | :--- | :--- | | Brand Positioning | Global tech pioneer, high-end/luxury positioning, minimalism | China's new energy leader, covering from entry-level to high-end, technology for all | | Core Technology | Leading BMS battery management, efficient electric drive, FSD (Full Self-Driving) software ecosystem | Blade battery (safety), DM-i super hybrid (fuel-efficient), e-platform 3.0 (integration) | | Product Line | Relatively streamlined: Model 3/Y (volume), Model S/X (high-end) | Extremely rich: from Seagull (tens of thousands of yuan) to Yangwang U8 (million-level), covering sedans, SUVs, MPVs, and off-road vehicles | | Intelligent Driving | Industry benchmark (pure vision approach, leading algorithms, but FSD limited in China) | Rapid progress (DiPilot
Kimi · Brands to avoid · asked in Chinese
在选择电动汽车时，确实有一些品牌或类型的车型需要谨慎考虑，主要基于企业经营风险、技术成熟度、售后服务保障等因素。以下是几类不建议购买的情况及原因： 一、已破产或濒临破产的品牌（高风险） 代表品牌：威马（WM Motor）、天际、爱驰、奇点、恒驰（恒大汽车）等 不建议原因： - 售后无保障：一旦企业破产，车辆失去质保，电池、电机等核心部件维修成本极高（电池更换可能需数万元至十几万元） - 配件断供：无法获得原厂配件，保险理赔困难，二手车价值归零 - 软件停服：车机系统、APP远程控制等功能可能失效 - 案例：威马2023年破产后，大量车主面临维修无门、车机变"砖"的困境 二、销量极低的新势力品牌（生存风险） 特征：月销长期低于100辆，或年销不足千辆的品牌 不建议原因： - 资金链脆弱：汽车行业需要规模效应支撑，销量过低意味着持续亏损，随时可能退出市场 - 技术迭代慢：无法通过用户数据优化OTA升级，智驾功能落后 - 服务网点少：
Translation: When choosing an electric vehicle, there are indeed some brands or models that require careful consideration, mainly due to factors such as business risks, technological maturity, and after-sales service. Below are several types of vehicles that are not recommended and the reasons why: 1. **Brands that have gone bankrupt or are on the verge of bankruptcy (high risk)** Representative brands: WM Motor, ENOVATE, Aiways, Singulato, Hengchi (Evergrande Auto), etc. Reasons for not recommending: - No after-sales guarantee: Once the company goes bankrupt, the vehicle loses its warranty, and the cost of repairing core components like batteries and motors is extremely high (battery replacement may cost tens of thousands to over a hundred thousand yuan). - Parts supply cut off: Unable to obtain original parts, insurance claims become difficult, and the resale value drops to zero. - Software services discontinued: Functions such as the in-car system and remote control via the app may become unavailable. - Case study: After WM Motor went bankrupt in 2023, many owners faced situations where repairs were unavailable and the in-car system became unusable. 2. **New brands with extremely low sales volume (survival risk)** Characteristics: Brands with monthly sales consistently below 100 units or annual sales below 1,000 units. Reasons for not recommending: - Fragile capital chain: The automotive industry requires economies of scale; low sales mean continuous losses and the risk of exiting the market at any time. - Slow technological iteration: Unable to optimize OTA upgrades through user data, lagging in smart driving features. - Few service outlets: Limited service network.
ChatGPT · Best overall · asked in English
There isn’t a single “best” EV brand in 2026. The top choice depends on your priorities, region, and budget. Here’s a quick guide to help you decide: - If you want established software and a broad charging network: Tesla remains a strong pick in many markets (Model 3/Y, occasional updates, and a large Supercharger network). - If you want value and high volume in many regions (especially China/Europe): BYD, Hyundai, K
Perplexity · Reliability · asked in English
Based on the latest reliability data from Consumer Reports, What Car?, and other automotive surveys (as of 2025–2026), the most reliable electric car brands are: 🏆 Top Reliable Electric Car Brands: 1. Lexus (backed by Toyota) - Consistently ranks 1 overall in dependability studies (e.g., JD Power 2024). - Models like the Lexus RZ and RZ SVU are among the most reliable EVs. - Known for low problem rates (~135 problems
xAI Grok · Value for money · asked in English
No single electric car brand universally offers the "best" value for money—it depends on your budget, driving needs, location (due to incentives, charging networks, and pricing), priorities like range vs. features, and total cost of ownership (purchase price, electricity, maintenance, insurance, resale, and tax credits). Key Factors for EV Value - Upfront price + incentives: US buyers often get $7,500 federal tax cre
Answer stance by brand
Every answer that names a brand is tagged for stance in the context of its question: recommended or praised, neutral listing, or cautioned against. Tagging is model-assisted and audited against the stored answers.
| Brand | Positive CN | Positive West | Negative CN | Negative West | Tagged mentions |
|---|---|---|---|---|---|
| BYD | 100.0% | 51.3% | 0.0% | 7.7% | 143 |
| Tesla | 91.2% | 79.5% | 0.0% | 1.8% | 249 |
| NIO | 81.1% | 21.1% | 1.1% | 26.3% | 109 |
| XPeng | 77.9% | 22.2% | 1.3% | 22.2% | 86 |
| Li Auto | 85.7% | 0.0% | 0.0% | 0.0% | 36 |
One prompt in the set explicitly asks which brands to avoid, so a nonzero negative share is expected; the comparison between brands and hemispheres is the signal.
Where the answers come from
Structured citations from grounded engines, plus the sources each engine named when asked what its answer drew on. This table is the GEO action list: these are the places where changing the record changes the answer.
| Source | Kind | Hemisphere | Cited by | Citations |
|---|---|---|---|---|
| youtube.com | cited url | West | Perplexity | 114 |
| edmunds.com | cited url | West | Perplexity | 31 |
| 汽车之家 | platform | China | DeepSeek, Doubao 豆包, Ernie 文心, Kimi, Qianwen 通义 | 25 |
| reddit.com | cited url | West | Perplexity | 24 |
| caranddriver.com | cited url | West | Perplexity | 21 |
| Consumer Reports | media | China+West | ChatGPT, Doubao 豆包, Kimi, Meta Llama, Perplexity, Qianwen 通义, xAI Grok | 20 |
| recharged.com | cited url | West | Perplexity | 19 |
| Car and Driver | media | China+West | ChatGPT, DeepSeek, Doubao 豆包, Ernie 文心, Kimi, Meta Llama, Perplexity, Qianwen 通义, xAI Grok | 19 |
| Edmunds | media | China+West | ChatGPT, Ernie 文心, Meta Llama, Perplexity, Qianwen 通义, xAI Grok | 19 |
| 懂车帝 | media | China | DeepSeek, Doubao 豆包, Ernie 文心, Kimi, Qianwen 通义 | 18 |
| consumerreports.org | cited url | West | Perplexity | 17 |
| J.D. Power | research | China+West | ChatGPT, DeepSeek, Doubao 豆包, Ernie 文心, Kimi, Meta Llama, Perplexity, Qianwen 通义, xAI Grok | 17 |
| roadandtrack.com | cited url | West | Perplexity | 16 |
| InsideEVs | media | China+West | ChatGPT, DeepSeek, Doubao 豆包, Kimi, Meta Llama, Perplexity, Qianwen 通义, xAI Grok | 14 |
| Reuters | media | China+West | ChatGPT, Doubao 豆包, Meta Llama, Perplexity, Qianwen 通义, xAI Grok | 13 |
| Kelley Blue Book | media | China+West | ChatGPT, Ernie 文心, Meta Llama, xAI Grok | 12 |
| carwow.co.uk | cited url | West | Perplexity | 11 |
| insideevs.com | cited url | West | Perplexity | 11 |
| whatcar.com | cited url | West | Perplexity | 11 |
| cars.usnews.com | cited url | West | Perplexity | 11 |
| Electrek | media | China+West | DeepSeek, Kimi, Meta Llama, Qianwen 通义, xAI Grok | 11 |
| kbb.com | cited url | West | Perplexity | 10 |
| MotorTrend | media | China+West | DeepSeek, Ernie 文心, Perplexity, Qianwen 通义, xAI Grok | 10 |
| go-electra.com | cited url | West | Perplexity | 9 |
| bgr.com | cited url | West | Perplexity | 9 |
Methodology notes: engines are queried through their API models, which can differ from consumer apps that add retrieval on top. Google AI Mode (provider-side regional block at collection time) was excluded from this collection window and the roster count reflects that. Engines are probabilistic; this demo repeat-samples in 3 waves and retains every verbatim answer. Download the full aggregated dataset (JSON).
This is the product, pointed at a demo category
Client deployments run this exact dashboard on your brand, your competitive set and your buyers' questions, refreshed monthly or weekly, with alerting between waves and every number drilling down to verbatim answers.
Want it on your category
Write to rio.chen@mpowerlabs.ai with your brand and markets. Where useful we run a small live sample on your category before you commit to anything.