EyeQ
EyeQ Kit is a software development kit (SDK) from Mobileye that gives automakers direct access to the EyeQ system-on-chip (SoC) for building custom in-vehicle applications.
Publisher review
EyeQ Kit is a software development kit (SDK) from Mobileye that gives automakers direct access to the EyeQ system-on-chip (SoC) for building custom in-vehicle applications. Targeted at OEMs developing advanced driver-assistance systems (ADAS) and autonomous driving (AD) features, the kit allows them to differentiate their user interfaces while relying on Mobileye's proven computer-vision stack. The kit is designed for teams that want to create visualization, augmented-reality head-up displays (AR-HUD), driver monitoring systems (DMS), human-machine interfaces (HMI), and in-vehicle infotainment (IVI) applications without building perception from scratch. It is already deployed in over 125 million vehicles, giving it a massive production-proven base.
EyeQ Kit runs on the EyeQ 6 family of SoCs, which are purpose-built for automotive-grade workloads. The EyeQ 6 High delivers 34 deep-learning TOPS (int8) on a 7nm process and can be clustered: two chips power Mobileye SuperVision (eyes-on, hands-off), three chips power Mobileye Chauffeur (eyes-off, hands-off), and four chips power Mobileye Drive (no driver). The EyeQ 6 Lite offers 5 DL TOPS for front-camera ADAS. The SDK provides an OpenCL runtime environment, TensorFlow support, a standard Linux environment for third-party middleware, and an x86-based development platform that covers the full cycle from bring-up to performance tuning. Automakers can consolidate other electronic control units (ECUs) into the Mobileye system to reduce integration complexity and cost.
EyeQ Kit competes directly with NVIDIA Drive Orin and Qualcomm Snapdragon Ride. Compared to NVIDIA Drive Orin, which offers up to 254 TOPS but at higher power consumption and cost, the EyeQ 6 High's 34 TOPS per chip is lower in raw compute but optimized for vision-specific tasks, delivering better performance per watt for typical ADAS workloads. Qualcomm Snapdragon Ride targets similar ADAS/AV tiers with flexible AI accelerators, but Mobileye's advantage lies in its integrated REM crowdsourced mapping and RSS (Responsibility-Sensitive Safety) driving policy, which are pre-validated and reduce OEM development risk. Analysts note that Mobileye offers lower cost, better performance for vision-centric tasks, and faster time-to-market, especially for automakers scaling from basic ADAS to full autonomy using a modular chip portfolio.
The honest trade-offs: First, EyeQ Kit ties OEMs to Mobileye's core technologies, creating dependency on a single supplier for perception and mapping. Second, the EyeQ 6 High's 34 TOPS is significantly lower than competitors like NVIDIA Orin (254 TOPS), which may limit headroom for compute-heavy sensor fusion or transformer-based models. Third, Mobileye faces pricing pressure in China, where local competitors offer lower-cost alternatives. Fourth, while the SDK is open for application development, the underlying computer-vision algorithms and driving policy remain proprietary, limiting full customization of the perception stack.
How it works
-
OEM application development
Enables automakers to build and deploy custom visualization, AR-HUD, DMS, HMI, and IVI applications on EyeQ SoCs.
-
OpenCL runtime environment
Provides an industry-standard OpenCL runtime for parallel computing, allowing efficient use of EyeQ's hardware accelerators.
-
TensorFlow support
Integrates TensorFlow for deploying neural network models, leveraging Mobileye's pre-trained vision models.
-
Standard Linux environment
Runs a standard Linux OS, enabling third-party middleware and libraries to be used alongside Mobileye's stack.
-
x86 development platform
Offers an x86-based platform for the full development cycle, from functional bring-up to performance tuning.
-
ECU consolidation
Allows consolidation of other electronic control units into the EyeQ system, reducing integration complexity and cost.
-
Access to Mobileye core technologies
Provides access to industry-leading computer vision, REM crowdsourced mapping, and RSS-based driving policy.
Strengths and trade-offs
Strengths
- Deployed in over 125 million vehicles, providing unmatched production validation and reliability data.
- EyeQ 6 High delivers 34 DL TOPS (int8) at 7nm, optimized for vision tasks with lower power consumption than general-purpose GPUs.
- Integrated REM crowdsourced mapping and RSS driving policy reduce OEM development risk and accelerate time-to-market.
- Modular chip scaling from EyeQ 6 Lite (5 TOPS) to clusters of four EyeQ 6 High chips enables a single platform from ADAS to full autonomy.
Trade-offs
- Dependency on Mobileye's proprietary perception and mapping technologies limits OEMs' ability to fully customize the core driving stack.
- EyeQ 6 High's 34 TOPS is significantly lower than NVIDIA Drive Orin's 254 TOPS, potentially limiting compute for sensor fusion or transformer models.
- Faces pricing pressure in China from local competitors offering lower-cost ADAS solutions.
- The SDK is open for application development but does not expose the underlying computer-vision algorithms for modification.
Pricing context
Pricing is not publicly disclosed; Mobileye typically negotiates per-vehicle licensing fees with OEMs, with costs varying by chip tier (EyeQ 6 Lite vs. EyeQ 6 High) and volume.
Getting started with EyeQ
-
Request access to EyeQ Kit
Contact Mobileye's OEM sales team to request access to the EyeQ Kit SDK. Provide your company details, target application (e.g., ADAS, HMI), and expected volume. Sign a non-disclosure agreement to receive the SDK package.
-
Set up x86 development environment
Install the EyeQ Kit SDK on an x86-based development machine. Follow the provided documentation to configure the OpenCL runtime and TensorFlow integration. Ensure your system meets the minimum hardware requirements for simulation.
-
Load pre-trained vision models
Import Mobileye's pre-trained computer vision models into your project using the SDK's TensorFlow support. These models handle object detection, lane recognition, and traffic sign classification, allowing you to focus on application logic.
-
Build a custom HMI prototype
Use the SDK's APIs to create a human-machine interface that visualizes perception outputs. For example, overlay detected objects on a camera feed or design an AR-HUD display. Test the prototype on the x86 platform using recorded vehicle data.
-
Deploy to EyeQ 6 SoC
Compile your application for the EyeQ 6 family of SoCs using the provided cross-compilation tools. Flash the binary onto an EyeQ 6 evaluation board and verify functionality with real sensor inputs. Tune performance using the SDK's profiling tools.
Frequently Asked Questions
What is EyeQ Kit and what does it do?
EyeQ Kit is a software development kit from Mobileye that gives automakers direct access to the EyeQ system-on-chip for building custom in-vehicle applications. It targets OEMs developing ADAS and autonomous driving features, allowing them to differentiate user interfaces while using Mobileye's computer-vision stack.
Which EyeQ chips does the EyeQ Kit support?
EyeQ Kit runs on the EyeQ 6 family of SoCs, including the EyeQ 6 High with 34 deep-learning TOPS on a 7nm process and the EyeQ 6 Lite with 5 DL TOPS. Multiple EyeQ 6 High chips can be clustered for higher autonomy levels like Mobileye SuperVision or Chauffeur.
How does EyeQ Kit compare to NVIDIA Drive Orin?
EyeQ Kit's EyeQ 6 High offers 34 TOPS versus NVIDIA Drive Orin's 254 TOPS, but it is optimized for vision-specific tasks with better performance per watt for typical ADAS workloads. Mobileye also provides integrated REM mapping and RSS driving policy, reducing OEM development risk and cost.
What are the main trade-offs of using EyeQ Kit?
EyeQ Kit ties OEMs to Mobileye's proprietary perception and mapping technologies, creating supplier dependency. The EyeQ 6 High's 34 TOPS is lower than competitors, potentially limiting compute for heavy sensor fusion. The SDK does not expose underlying computer-vision algorithms for full customization.
What applications can automakers build with EyeQ Kit?
Automakers can build custom visualization, augmented-reality head-up displays, driver monitoring systems, human-machine interfaces, and in-vehicle infotainment applications. The kit provides an OpenCL runtime, TensorFlow support, and a standard Linux environment for third-party middleware and development.
How much does EyeQ Kit cost and how is it licensed?
Pricing is not publicly disclosed. Mobileye typically negotiates per-vehicle licensing fees with OEMs, with costs varying by chip tier, such as EyeQ 6 Lite versus EyeQ 6 High, and volume. The kit is already deployed in over 125 million vehicles globally.
Alternatives
How EyeQ compares
Direct head-to-head against 2 competitors. Picked by 7wData.
EyeQ
- Pricing
- Pricing is not publicly disclosed; Mobileye typically negotiates per-vehicle licensing fees with OEMs, with costs varying by chip tier (EyeQ 6 Lite vs. EyeQ 6 High) and volume.
- Target
- EyeQ Kit is a software development kit (SDK) from Mobileye that gives automakers direct access to the EyeQ system-on-chip (SoC) for building custom in-vehicle applications.
- Strength
- Deployed in over 125 million vehicles, providing unmatched production validation and reliability data.
- Watch for
- Dependency on Mobileye's proprietary perception and mapping technologies limits OEMs' ability to fully customize the core driving stack.
Canary Technologies
- Pricing
- $300/mo
- Target
- Hotels seeking contactless check-in solutions
- Deployment
- Cloud-based
- Strength
- Streamlined guest management and check-in process
- Watch for
- Limited customization options
Duve
- Pricing
- $900/mo
- Target
- Hotels focusing on guest experience enhancement
- Deployment
- Cloud-based
- Strength
- Comprehensive guest experience platform
- Watch for
- Higher pricing compared to competitors
User reviews
No user reviews yet. Be the first to write one.
Sources
Reporting on this tool draws on these publicly available sources.