NVIDIA releases open AI models spanning robotics to healthcare

NVIDIA Corp. (NASDAQ: NVDA) released a collection of open AI models, datasets and development tools across multiple industries including autonomous vehicles, robotics, healthcare and enterprise applications, according to a company statement.

The release includes models from the Nemotron family for enterprise AI applications, the Cosmos platform for physical AI systems, the new Alpamayo family for autonomous vehicle development, Isaac GR00T for robotics and Clara models for biomedical applications.

The company made available training frameworks and datasets including 10 trillion language training tokens, 500,000 robotics trajectories, 455,000 protein structures and 100 terabytes of vehicle sensor data.

Technology companies including Bosch, CrowdStrike, Cohesity, Fortinet, Franka Robotics, Palantir, Salesforce, ServiceNow, Hitachi and Uber are implementing these open model technologies in their products and services.

The Nemotron family includes speech recognition models, multimodal retrieval systems and safety models. Bosch is using Nemotron Speech for vehicle interaction systems, while ServiceNow incorporates the technology in its Apriel model family. CrowdStrike, Cohesity and Fortinet are adopting Nemotron Safety models for their AI applications.

The Cosmos platform offers world foundation models for physical AI development. Cosmos Reason 2 provides vision-language capabilities for robots and AI agents, while Cosmos Transfer 2.5 and Cosmos Predict 2.5 generate synthetic videos across different environments.

The Alpamayo family targets autonomous vehicle development with reasoning-based models and simulation tools. It includes Alpamayo 1, described as the first open large-scale reasoning model for autonomous vehicles, and AlpaSim, a simulation framework for training and evaluation.

Clara AI models for healthcare include La-Proteina for protein design, ReaSyn v2 for drug synthesis planning, KERMT for computational safety testing and RNAPro for RNA molecule structure prediction.

You May Also Be Interested In