NVIDIA Accelerates AI Innovation with New Open Models and Tools

NVIDIA has unveiled a suite of new open models, datasets, and tools designed to drive AI advancements across various industries. This release includes expansions to the NVIDIA Nemotron family for speech, multimodal, and safety applications, as well as the introduction of the NVIDIA Alpamayo family for autonomous vehicle development. The launch also features updates to the NVIDIA Cosmos platform for physical AI and the NVIDIA Clara suite for biomedical research, providing developers with unprecedented resources to innovate in language processing, robotics, scientific research, and autonomous systems.

The Nemotron family now includes models for speech recognition, multimodal retrieval-augmented generation (RAG), and safety, with the Nemotron Speech model delivering real-time, low-latency speech recognition for live captions and AI applications. The Nemotron RAG models enhance document search and information retrieval with highly accurate multilingual and multimodal data insights. Safety models, such as the Llama Nemotron Content Safety model and Nemotron PII, strengthen the trustworthiness of AI applications by detecting sensitive data with high accuracy.

In the autonomous vehicle sector, the NVIDIA Alpamayo family introduces Alpamayo 1, the first open, large-scale reasoning VLA model for autonomous vehicles, enabling vehicles to understand their surroundings and explain their actions. The AlpaSim simulation framework supports closed-loop training and evaluation of reasoning-based AV models across diverse environments and edge cases. Additionally, NVIDIA has released over 1,700 hours of driving data as part of its Physical AI Open Datasets, covering rare and complex real-world scenarios essential for advancing reasoning architectures.

Expanding AI Capabilities in Biomedical Research

NVIDIA Clara for Healthcare and Life Sciences introduces new AI models aimed at lowering costs and accelerating treatment development. La-Proteina enables the design of large, atom-level-precise proteins for research and drug candidate development, providing scientists with tools to study diseases previously considered untreatable. ReaSyn v2 ensures AI-designed drugs are practical to synthesize by incorporating a manufacturing blueprint into the discovery process. KERMT provides high-accuracy computational safety testing early in development, predicting how potential drugs will interact with the human body. RNAProunlocks the potential of personalized medicine by predicting the complex 3D shapes of RNA molecules. These advancements are supported by an NVIDIA dataset of 455,000 synthetic protein structures, helping researchers build more accurate AI models.

Empowering Developers with Open Resources

NVIDIA's commitment to open-source innovation is evident in its contribution of training frameworks and one of the world’s largest collections of open multimodal data, including 10 trillion language training tokens, 500,000 robotics trajectories, 455,000 protein structures, and 100 terabytes of vehicle sensor data. This scale of diverse open resources is designed to accelerate innovation in language processing, robotics, scientific research, and autonomous vehicles. Leading technology companies, including Bosch, CodeRabbit, CrowdStrike, Cohesity, Fortinet, Franka Robotics, Humanoid, Palantir, Salesforce, ServiceNow, Hitachi, and Uber, are adopting and building on NVIDIA’s open model technologies.

The open models, data, and frameworks are now available on GitHub, Hugging Face, and various cloud and AI infrastructure platforms, providing developers with flexible access to supporting resources. Many of these models are also available as NVIDIA NIM microservices for secure, scalable deployment on any NVIDIA-accelerated infrastructure. NVIDIA's latest releases signal a significant push to democratize AI development, empowering developers and researchers to push the boundaries of what is possible in AI innovation.