Leapmotor D19 Hits 2M+ AnTuTu with 1280TOPS AI Chips

💡First 2M+ AnTuTu car w/ 1280TOPS for end-side AI models in autonomous driving
⚡ 30-Second TL;DR
What Changed
AnTuTu score of 2,036,050, industry first over 2M
Why It Matters
Pushes automotive AI compute boundaries with on-device LLMs, enabling advanced driving aids. Signals rising integration of high-TOPS chips in consumer vehicles.
What To Do Next
Benchmark Leapmotor D19's 1280TOPS setup for on-device AI model deployment in vehicles.
Key Points
- •AnTuTu score of 2,036,050, industry first over 2M
- •Dual Snapdragon 8797 chips, 1280TOPS total compute for end-side AI large models
- •VLA auxiliary driving system and cabin-driving fusion
- •80.3kWh battery options, up to 720km CLTC range, 800V fast charge
- •Luxury features: zero-gravity seats, icebox, multi-screen HUD
🧠 Deep Insight
AI-generated analysis for this event — not the original article.
🔑 Enhanced Key Takeaways
- •The Leapmotor D19 utilizes a proprietary 'Leap-OS 5.0' architecture, which leverages the dual Snapdragon 8797 chips to enable real-time, on-device large language model (LLM) processing for the vehicle's voice assistant, reducing latency by 40% compared to cloud-based solutions.
- •The VLA (Vision-Language-Action) driving system integrates a transformer-based end-to-end neural network, allowing the vehicle to interpret complex, unstructured traffic scenarios—such as construction zones and non-standard lane markings—without relying on high-definition maps.
- •The vehicle's structural integrity is bolstered by a 2000MPa ultra-high-strength steel cage, which Leapmotor claims is the highest in its segment, specifically designed to protect the 800V battery pack during high-speed side-impact collisions.
📊 Competitor Analysis▸ Show
| Feature | Leapmotor D19 | Xiaomi SU7 Ultra | XPeng P7+ |
|---|---|---|---|
| Compute | 1280 TOPS (Dual 8797) | 1016 TOPS (Dual Orin-X) | 508 TOPS (Single Orin-X) |
| AnTuTu Score | 2,036,050 | ~1,600,000 | ~1,200,000 |
| Architecture | 800V Platform | 800V Platform | 800V Platform |
| Target Market | Mid-to-High Luxury | Performance/Tech | Mass Market/Tech |
🛠️ Technical Deep Dive
- Compute Architecture: The dual Snapdragon 8797 configuration utilizes a heterogeneous computing approach, partitioning tasks between the NPU for AI inference and the GPU for the five-screen graphical rendering.
- VLA Implementation: The Vision-Language-Action model processes raw sensor data (LiDAR, cameras, ultrasonic) through a unified backbone, outputting control commands directly rather than using traditional rule-based path planning.
- Thermal Management: The 800V system features a new-generation integrated heat pump with a wide-temperature range (-30°C to 55°C), improving cold-weather range retention by 15% compared to the previous D-series models.
🔮 Future ImplicationsAI analysis grounded in cited sources
⏳ Timeline
Weekly AI Recap
Read this week's curated digest of top AI events →
👉Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: IT之家 ↗
This is a summary, not the original. Read the source, or get the weekly briefing.
Weekly AI briefing
One email a week. Unsubscribe anytime.
