NVIDIA Launches Alpamayo Open AI Models To Advance Safe Autonomous Driving

(RTTNews) - NVIDIA has unveiled the Alpamayo family of open-source AI models, simulation tools, and datasets at CES 2026, designed to accelerate safe, reasoning-based autonomous vehicle (AV) development.

The Alpamayo portfolio includes Alpamayo 1, AlpaSim, and Physical AI Open Datasets, enabling vehicles to perceive, reason, and act with humanlike judgment. These resources allow developers to fine-tune, distill, and test models that enhance safety, robustness, and scalability. Mobility leaders such as JLR, Lucid, and Uber, along with AV research groups like Berkeley DeepDrive, are expected to leverage Alpamayo to fast-track level 4 deployment roadmaps.

Autonomous vehicles must operate safely across a vast range of driving conditions, including rare and complex scenarios known as the "long tail." Traditional AV architectures often separate perception and planning, limiting scalability in unusual situations. While end-to-end learning has advanced capabilities, tackling long-tail challenges requires models that can reason about cause and effect beyond their training data.

The Alpamayo family introduces reasoning-based vision language action (VLA) models that apply chain-of-thought logic to decision-making. By thinking through novel scenarios step by step, these models improve driving performance and explainability — a critical factor in building trust and safety in intelligent vehicles. All Alpamayo systems are supported by the NVIDIA Halos safety framework, ensuring reliable and transparent deployment.