NVIDIA has released an open-source reasoning AI model for autonomous driving, writes Nick Flaherty.

The Alpamayo 1 open-reasoning AI model works with a new tool called AlpaSim and Physical AI Open Datasets. This enables the development of vehicles that allow developers to optimise and test models for greater safety, robustness and scalability.

Car makers JLR and Lucid, and researchers Berkeley DeepDrive, are using the Alpamayo 1 model for deployment of reasoning-based systems at Level 4 of the SAE autonomous driving scale.

The Alpamayo architecture tackles rare, complex scenarios by combining perception and planning to safely reason about cause and effect, especially when situations fall outside a model’s training experience. The architecture integrates three foundational pillars: open models, simulation frameworks and datasets.

The family of models introduces chain-of-thought, reasoning-based vision language action (VLA) models that can think through novel or rare scenarios step-by-step, improving driving capability and explainability.

“The ChatGPT moment for physical AI is here – when machines begin to understand, reason and act in the real world,” said Jensen Huang, founder and CEO of NVIDIA. “Robotaxis are among the first to benefit. Alpamayo brings reasoning to autonomous vehicles, allowing them to think through rare scenarios, drive safely in complex environments and explain their driving decisions.”

Developers can adapt Alpamayo 1 into smaller runtime models for vehicle development, or use it as a basis for AV development tools such as reasoning-based evaluators and auto-labelling systems.