Mobileye EyeQ6 High
The Mobileye EyeQ6 High (EyeQ6H) is a high-performance system-on-chip designed for premium advanced driver-assistance systems (ADAS) and partial autonomous driving, delivering 34 deep-learning TOPS on a 7nm process to handle multi-sensor perception and fusion in a single device.
- Function
- Vision + sensor fusion + REM mapping
- Process Node
- 7nm
- Target Level
- L2+ / L2++ (scales to L3/L4 paired)
- AI Performance
- 34 DL TOPS (INT8)
Where to buy in Singapore
Automotive-grade sensing and compute is supplied to vehicle manufacturers under design-win programmes, in production volumes and with multi-year validation — it is not sold as individual units through resellers. Approach the manufacturer directly if you are specifying for a production programme.
We keep this page because the comparison is still useful.
Highlights
- ✓Single-chip integration of up to 11 sensors
- ✓High compute efficiency (34 TOPS at low power)
- ✓Proven EyeQ platform shipped at massive scale
- ✓Scales from L2+ to L4 by pairing chips
Overview
Overview
The Mobileye EyeQ6 High is a centralized system-on-chip (SoC) purpose-built for premium driver-assistance and partial autonomous driving applications. It consolidates computer vision processing, multi-sensor fusion, and Mobileye's REM (Road Experience Management) crowdsourced mapping into a single piece of silicon — an architectural approach that reduces system complexity compared to multi-chip designs.
How It Works
At its core, the EyeQ6H runs deep-learning inference workloads that interpret data from the vehicle's sensor array in real time. A single chip can ingest and process inputs from up to 11 cameras and radars simultaneously, enabling the perception pipelines needed for lane-keeping, object detection, path planning, and situational awareness tasks associated with L2+ and partial-autonomy driving.
The chip is manufactured on a 7nm process node, which enables high computational density within a constrained power envelope. It delivers 34 deep-learning TOPS — roughly triple the compute of the prior-generation EyeQ5H — while consuming only about 25% more power, reflecting the efficiency gains of the newer process node.
Scalability and Deployment
The EyeQ6H is the compute foundation for several of Mobileye's own driving software stacks:
- Mobileye SuperVision — a hands-on L2+ system using surround-view cameras for highway and urban driving assistance
- Mobileye Chauffeur — targeting hands-free L2++ operation, supported by pairing two EyeQ6H chips for increased compute headroom
- Mobileye Drive — Mobileye's platform aimed at L3 and L4 autonomy levels
Pairing two EyeQ6H chips in a dual configuration is the intended path for next-generation hands-free SuperVision deployments and can scale toward L3 and L4 system architectures.
Who Uses It and Why
The EyeQ6H is targeted at automotive OEMs and Tier-1 suppliers integrating advanced driver-assistance features into production vehicles. Mobileye has a long track record of shipping EyeQ-series silicon across tens of millions of vehicles, and the EyeQ6H is in production availability, making it a proven compute platform rather than a research prototype. Engineers and system architects working on centralized ADAS domain controllers use the EyeQ6H for its combination of high inference throughput, multi-sensor input capacity, and compatibility with Mobileye's broader software and mapping ecosystem.
Specifications
| Function | Vision + sensor fusion + REM mapping |
| Process Node | 7nm |
| Target Level | L2+ / L2++ (scales to L3/L4 paired) |
| AI Performance | 34 DL TOPS (INT8) |
| Sensor Capacity | Up to 11 cameras/radars per chip |
| Compute vs EyeQ5H | about 3x compute, ~25% more power |