JAKARTA NVIDIA announced a new Artificial Intelligence (AI) model called Alpamayo-R1. This technology was developed to build physical AI devices such as robots or autonomous vehicles.
Introduced at the AI NeurIPS conference, the Alpamayo-R1 is the first visual language action model to focus on autonomous driving. This visual language model is able to process text and images simultaneously.
With this capability, autonomous vehicles can see and recognize the surrounding environment well. That way, autonomous vehicles can make more natural driving decisions, similar to humans.
The Alpamayo-R1 is a development of the NVIDIA Cosmos-Readon model, a reasoning model that takes into account decisions deeply before responding. The family of Cosmos models was originally released in January 2025.
SEE ALSO:
Through the launch of this new model, NVIDIA hopes that its reasoning model can provide the necessary 'healthy' autonomous vehicle. This ability is important for more complex driving decisions.
Along with this new vision model, NVIDIA also launched a series of resources called Cosmos Cookbooks. These resources include step-by-step guidance and a new post-training workflow on GitHub.
This guide aims to help developers to better use and train Cosmos models. The Cookbook includes data curation, synthetic data creation, model evaluation, and is available on GitHub and Hugging Face.
The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)