Tier IV Releases Open-Source Autonomous Driving AI Models
The models, built for Autoware 2.0, range from camera-only systems to setups combining LiDAR, radar and high-definition maps.
Updated on Nov. 11, 2026 in Artificial Intelligence
Live Poll
Would you trust a self-driving car to drive you around?
Tier IV has released open-source end-to-end autonomous driving AI models through Autoware. The models span configurations from camera-only setups to systems combining LiDAR, radar and high-definition maps.
Why it matters
Tier IV's view is that automakers and suppliers should be able to adopt AI without relying on one model or chip platform. The release gives developers a choice of models and tools to adapt them into vehicles.
The released models range from 16 million parameters for Tier IV's Meteor to 79.9 million for AutoE2E, with Nvidia's Alpamayo 2 Super at 34 billion. Planning horizons span four seconds for LatentDrive to six seconds for OnePlanner.
The players
Tier IV
A Japanese autonomous driving technology company that develops the Autoware open-source software ecosystem and the models described here.
Autoware Foundation
An open-source organization that stewards the Autoware autonomous driving software, with which AutoE2E was co-developed.
Nvidia
A semiconductor and AI computing company whose Alpamayo 2 Edge model was integrated with Autoware.
The details
Developers can select a model to suit their performance targets, and Autoware supplies tools to adapt and integrate it into vehicles. AutoE2E was co-developed with the Autoware Foundation and runs beyond Nvidia hardware, while Nvidia's camera-only Alpamayo 2 Edge has been integrated with Autoware.
Timeline
• Article on the open-source release published November 11, 2026.
The Tech Race
The models are built for Autoware 2.0, an architecture pairing AI with rule-based software, extending that framework to end-to-end learning. The release positions open-source autonomous driving stacks against proprietary, single-vendor platforms.
Automakers and suppliers can choose among models of different sizes and sensor setups rather than committing to one chip platform. Drivers would see the effects only as these systems are adopted in production vehicles.
The takeaway
Open, modular autonomous driving stacks let developers match model size and sensor setup to their own hardware and performance targets. That flexibility could lower barriers for companies seeking to adopt AI-driven driving without locking into a single vendor.
Further reading
Explore more Artificial Intelligence coverage to see how open models are reshaping autonomous systems.
Source note: This article includes information reported by Automotive World.
Live Poll
Would you trust a self-driving car to drive you around?
