TIER IV announced on November 11, 2026, the development and open-source availability of five end-to-end autonomous-driving AI models for Autoware, spanning camera-only systems and configurations that add LiDAR, radar and high-definition maps. The lineup gives automakers and technology partners several options for optimizing models and integrating them into vehicle systems.
The lineup comprises NVIDIA Alpamayo 2 Edge, AutoE2E, METEOR, OnePlanner and LatentDrive. Their end-to-end designs connect perception of the environment with driving-trajectory generation, but the models differ in sensor requirements, planning approach and readiness: some are integrated with Autoware, while vehicle adaptation or integration continues for others.
TIER IV presents the models as concrete technological outcomes for Autoware 2.0, the next-generation architecture it detailed on November 2, 2026. That architecture combines AI-based and conventional rule-based approaches, adds a flexible indexing scheme for modular software integration and provides a Reference Design framework intended to facilitate commercial adoption across vehicles, applications and operating environments.
Different Models for Different Vehicles
TIER IV reports that Alpamayo 2 Edge has 2 billion parameters, compared with 34 billion in Alpamayo 2 Super, and uses only camera images and reasoning to generate driving trajectories. Its Autoware integration is complete, but TIER IV is still optimizing the model and adapting it to real vehicles.
The Autoware Foundation and TIER IV lead AutoE2E’s development. TIER IV puts it at 79.9 million parameters and says it can run on computing platforms beyond NVIDIA hardware, with configurations ranging from cameras alone to systems incorporating LiDAR and radar. Its fast–slow architecture pairs a Reactive Model for real-time decisions with World and Reasoning Models that predict environmental changes and support driving decisions.
METEOR, developed independently by TIER IV and aimed primarily at privately owned passenger vehicles, has 16 million parameters. TIER IV says its convolutional neural network uses a bird’s-eye-view representation to handle perception through trajectory generation in one network without relying on HD maps. It accepts images from one to eight cameras, can add LiDAR and radar, and supports NVIDIA and Renesas semiconductor platforms.
TIER IV says OnePlanner combines perception and motion planning. It currently uses BEVFusion to create a bird’s-eye-view representation from LiDAR point clouds and passes features extracted during perception directly to the planner, reducing intermediate conversions into object positions and velocities. It also incorporates HD maps, route information, traffic-signal states and vehicle motion states to generate six-second trajectories. Autoware integration is complete.
LatentDrive uses V-JEPA, a self-supervised learning approach, to represent features from camera video in a latent space and generate trajectories without intermediate object detection and tracking. TIER IV says it needs no HD map and produces four-second trajectories using video, route information and vehicle motion states. It supports one or three cameras, while Autoware integration and in-vehicle validation remain underway.
What AutoE2E V1.0 Includes
The Autoware Foundation’s Robotaxi Working Group, sponsored by TIER IV, released AutoE2E v1.0 on November 11, 2026, under the Apache 2.0 license for research and commercial use, with trained checkpoints also available. The release includes the Reactive Model, which runs at 10 Hz, fuses six camera views into a bird’s-eye representation and plans 6.4 seconds of acceleration and steering. The world-model and reasoning layers are still being developed to work with the same planner.
The foundation reports that the fine-tuned AutoE2E model achieved an average displacement error of 0.744 meters over three seconds in recorded-data validation, ranking second among eight open-weight configurations in its comparison. In that open-loop test, models predicted from logged camera frames and were scored against the human driver’s actual path; AutoE2E also received a map raster and route, while inputs varied across models. The result therefore concerns trajectory prediction on recorded data, not driverless road performance.
AutoE2E is integrated with Autoware for open-loop evaluation, and real-vehicle closed-loop testing is underway to assess interactive driving. The foundation says deployment performance and resource requirements remain subject to platform-specific evaluation, with further on-road trials planned for 2027. The project targets applications from L2++ hands-free driving assistance to future L4 driverless mobility, but L4 remains a development goal rather than a current capability.
Commercialization and Development Workflow
For METEOR, TIER IV says it uses CoMET, the automated labeling platform within Co-MLOps, together with agentic AI to automate data labeling and model training, evaluation and improvement, eliminating the need for manual coding and labeling. Automakers and component suppliers can optimize the open-source model for their own vehicle specifications and performance requirements.
Founder and CEO Shinpei Kato said the multi-model strategy is intended to help manufacturers and suppliers “internalize technologies according to their applications and performance requirements, without depending on a particular AI model or semiconductor platform.”
NVIDIA automotive vice president Ali Kani said the Alpamayo 2 Edge integration gives developers an open foundation for moving the technology from research into their own vehicles.
TIER IV’s stated next steps are further performance improvements, in-vehicle optimization, commercialization with automakers and partners through Autoware, and expansion across vehicle types and applications. Autoware 2.0 supplies the shared architecture, but AutoE2E’s unfinished closed-loop testing shows that open-source availability and validated vehicle deployment remain separate milestones.
Credit: Source link
