About
EyeQ®
The System-on-Chip for Automotive Applications
Mobileye’s proven leadership is based on our EyeQ® System-on-Chip. More than 30 vehicle manufacturers have chosen EyeQ for its ability to support complex and computationally intense vision tasks, while meeting optimal, ambitious power-performance-cost targets.
EyeQ® is the brain behind everything Mobileye does – from driver-assistance to self-driving technologies. It’s a highly efficient, scalable, and proven SoC that is driving the mobility revolution.
20 Years,
100 Million Vehicles
on the Road
Efficiency through
Unique Architecture of
Diversified Accelerators
One Automotive-Grade
SoC Family, Many
Mobility Solutions
20 Years, 100 Million Vehicles on the Road
In over 20 years of technological advancement, EyeQ has emerged as the trusted solution for processing advanced mobility technologies.
100M+ EyeQs shipped
to date
700+ Models on the market
incorporate EyeQ
38+ Automakers place
their trust in EyeQ
7 Generations
of EyeQ
Industry Firsts for Over a Decade
VOLVO
EyeQ1
2007 Camera/Radar ACCGM, BMW
EyeQ2
2011 Camera-only FCWAUDI
EyeQ3
2014 Camera-only AEBNISSAN
EyeQ3
2016 Camera-only Hands-free DrivingHONDA
EyeQ4
2019 Industry First 100° CameraGEELY
EyeQ5
2021 Mobileye SuperVision™ LaunchVOLVO
EyeQ2
2010 Pedestrian AEBBMW
EyeQ2
2013 Camera-only ACCTESLA
EyeQ3
2015 Highway AutopilotBMW
EyeQ4
2018 REM™Mapping Harvesting Launch
BMW
EyeQ5
2021 Industry First 120° 8MP CameraEfficiency through Unique Architecture of Diversified Accelerators
By using highly efficient hardware accelerator architecture, EyeQ achieves state-of-the-art computer vision performance within a low-power envelope.
Heterogeneous Computing
Using the most suitable core for each task
From general-purpose CPU cores to high compute density accelerators, including deep learning neural networks.
Scalable Architecture
Utilizing the right mixture of accelerators to match the different EyeQ models’ needs.
Deep Learning Performance / Compute Density
XNN
Deep Learning Accelerator
Dedicated high-performance AI engine. The main source of horsepower for convolutional neural network.
PMA
Programmable Macro Array
A CGRA dataflow machine. Its unique architecture delivers outstanding performance for dense compute-vision and deep-learning algorithms that are unachievable in classic DSP architecture.
VMP
Vector Microcode Processor
A wide vector (VLIW and SIMD) machine with exceptional performance for short integral types common in computer-vision and deep-learning algorithms.
MPC
Multi-threaded Processor Cluster
More versatile than a GPU, and with higher efficiency than any CPU
CPU
General-Purpose Compute
Designing EyeQ’s hardware architecture with its intended software needs in mind, along with these heterogeneous accelerators, enables EyeQ to provide unparalleled performance, while requiring fewer TOPS and consuming less power.
UltimateCost-Performance Ratio
EyeQ in Action
Watch how the highly efficient, scalable, and proven technology of EyeQ enables our developmental self-driving vehicle to operate seamlessly in real-world traffic.
Full support for co-hosting third-party applications
The newest generations of EyeQ support open compute by offering a complete SDK package, and OpenCL environment and TensorFlow support, allowing customers to differentiate their solutions by deploying their algorithms on EyeQ. In addition, EyeQ supports co-hosting third-party applications, removing the need for additional processors and reducing the burden on the MCU.
One Automotive-Grade SoC, Many Mobility Solutions
Tailored specifically to deliver trusted mobility solutions, EyeQ is the only scalable automotive-grade SoC that can truly address the needs of both the driver-assist and autonomous-driving markets.
EyeQ Portfolio
One family of SoCs supports a full spectrum of mobility solutions: from a windshield-mounted, low-voltage front-camera driver-assist system, through a surround vision premium driver-assist solution, to processing of dozens of sensors, including high-resolution cameras, radars, and lidars for a full self-driving system.
