Nvidia launches Alpamayo, an open AI model that allows autonomous vehicles to ‘think like humans’

At CES 2026Nvidia launched Alpamayo, a new family of open source AI models, simulation tools, and datasets for training physical robots and vehicles designed to help autonomous vehicles reason through complex driving situations.

“ChatGPT’s moment for physical AI is here — when machines begin to understand, reason, and act in the real world,” Nvidia CEO Jensen Huang said in a statement. “Alpamayo brings reasoning to autonomous vehicles, allowing them to think through rare scenarios, drive safely in complex environments, and explain driving decisions.”

At the heart of Nvidia’s new family is Alpamayo 1, a 10 billion parameter chain-of-thought model, a vision language action model (VLA) based on reasoning that allows AVs to think more like humans to solve complex edge cases – like how to navigate a traffic light at a busy intersection – without prior experience.

“It does this by breaking down the problem into steps, thinking about all the possibilities, and then choosing the safest path,” said Ali Kani, Nvidia’s vice president of automotive, on Monday during a press conference.

Or as Huang said during his keynote on Monday: “Not only does (Alpamayo) take the sensor input and activate the steering, braking and acceleration, but also the reason for the action to be taken. It tells us what to do, the reasons for it. And then, of course, the trajectory.

The base code of Alpamayo 1 is available at Hugging Face. Developers can adapt Alpamayo to a smaller and faster version for vehicle development, use it to train simpler driving systems, or build devices on top like an automatic labeling system that automatically tags video data or an evaluator that checks if the car makes intelligent decisions.

“They can also use Cosmos to generate synthetic data and then train and test Alpamayo-based AV applications on a combination of real and synthetic datasets,” Kani said. Cosmos is Nvidia brand generative world modelAn AI system that creates a representation of the physical environment in order to make predictions and take action.

As part of the Alpamayo rollout, Nvidia also released an open dataset with more than 1,700 hours of driving data collected in various geographies and conditions, covering rare and complex real-world scenarios. The company also launched AlpaSim, an open-source simulation framework for validating autonomous driving systems. Available on GitHub, AlpaSim is designed to recreate real-world driving conditions, from sensors to traffic, so developers can safely test systems at scale.