Furthering its deep commitment to open source, NVIDIA is unveiling new open-source AI technologies for language, robotics and biology \u2014 contributing to an open ecosystem that broadens access to AI and fuels U.S. innovation.<\/p>\n
These open technologies will empower developers worldwide and strengthen economic growth through efficient reasoning, high-fidelity world generation and interactive physical AI systems accelerated on NVIDIA infrastructure.<\/p>\n
The new open models, data and tools are part of the NVIDIA Nemotron<\/a> family for AI reasoning, the NVIDIA Cosmos<\/a> platform for physical AI<\/a>, NVIDIA Isaac GR00T<\/a> for robotics and NVIDIA Clara<\/a> for biomedical AI.<\/p>\n NVIDIA is contributing these models, data and training frameworks to Hugging Face to make AI research and development more accessible. As a top contributor<\/a> to Hugging Face, with more than 650 open models and 250 open datasets now available, NVIDIA continues to expand access to cutting-edge AI resources for the global developer community.<\/p>\n \u201cOpen models are catalysts to AI innovation, making AI accessible, transparent and responsible,\u201d said Cl\u00e9ment Delangue, CEO of Hugging Face. \u201cNVIDIA\u2019s contributions to the open model ecosystem, commitment to open research for AI and Hugging Face\u2019s ecosystem will empower millions of developers to build advanced AI \u2014 together and in the open.\u201d<\/p>\n NVIDIA Nemotron Brings Ultra-Efficient Reasoning to Specialized AI Agents<\/p>\n AI agents use multiple models to see, retrieve, generate and reason. The latest open models in the NVIDIA Nemotron family unify these capabilities, enabling developers to build specialized, intelligent agents.<\/p>\n Nemotron Nano 3 uses a hybrid mixture-of-experts architecture to improve reasoning throughput for areas like software development, customer service and IT support. NVIDIA also released new open-source Nemotron datasets \u2014 including on multimodal training, multilingual personas<\/a> and privacy-preserving synthetic personal information<\/a> \u2014 for specialized model development. In addition, new NVIDIA NeMo tools, including NeMo Data Designer<\/a> for synthetic data generation and NeMo-RL<\/a> for advanced post-training and reinforcement learning, give developers greater control over model customization.<\/p>\n Leading Software Companies Building Agentic AI on Nemotron<\/p>\n Open-source models enable developers to build and adapt AI for their domains, transforming software into agentic systems that use tools, not just provide them. Leading software and services companies are building agentic software platforms using NVIDIA Nemotron.<\/p>\n Building on the success of the Apriel Nemotron model family that\u2019s post-trained with NVIDIA and ServiceNow-provided data, ServiceNow introduced its new Apriel 2.0 multimodal reasoning model that brings intelligence to cross-enterprise workflows for every industry, including regulated sectors like financial services, healthcare and telecom.\u00a0<\/p>\n \u201cOpen models are driving the next wave of enterprise transformation,\u201d said Joe Davis, executive vice president of platform engineering and AI at ServiceNow. \u201cApriel 2.0 represents one of the first open-weight multimodal reasoning models built for the enterprise \u2014 combining text, document and data understanding to power real-world workflows across industries. Together with NVIDIA, we\u2019re pairing reasoning transparency with performance efficiency, bringing explainable, secure and scalable AI to sectors where trust and compliance matter most.\u201d<\/p>\n Additional leading software companies are adopting NVIDIA\u2019s latest reasoning models to power their next-generation AI applications:<\/p>\n Palantir<\/a> is making Nemotron available via its Ontology<\/a> in Foundry<\/a> and AIP<\/a> to support an integrated technology stack with NVIDIA for operational AI. Cosmos and Isaac GR00T Open Models and Data for Physical AI and Robotics<\/p>\n To accelerate the training of robotic systems with humanlike reasoning and cognition, NVIDIA introduced major updates to its open models for physical AI, including Cosmos world foundation models and Isaac GR00T robot foundation models:<\/p>\n Cosmos Predict 2.5<\/a>: Unifies three models into one for rapid world simulation, generating 30-second videos from a single frame. NVIDIA also released the world\u2019s largest open-source dataset<\/a> for physical AI, now featuring 1,700 hours of multimodal driving sensor data from across the U.S. and Europe as well as GR00T training data, which has risen to the top 10 most-downloaded Hugging Face datasets of all time.<\/p>\n Leading companies \u2014 including Agility Robotics, Amazon Robotics, Figure AI, Skild AI, Milestone Systems and Uber \u2014 are adopting Cosmos or Isaac GR00T N models to generate synthetic data<\/a>, teach robots new behaviors and deploy real-world and physical AI agents at scale.<\/p>\n NVIDIA Clara Open Models for Healthcare and Life Sciences<\/p>\n New open models joining NVIDIA Clara, a family of models, tools and recipes built for accelerating scientific discovery, analyzing medical images and more, include:<\/p>\n Clara CodonFM: Learns the rules of RNA to reveal how changes in its code can improve the design of therapies and medicine. NVIDIA will contribute open models like CodonFM to the Chan Zuckerberg Initiative\u2019s virtual cells platform, accelerating open-source collaboration and model evaluation. Get Started With NVIDIA Open Models<\/p>\n Select NVIDIA Nemotron and Cosmos open models, trained on NVIDIA DGX Cloud, are available on build.nvidia.com<\/a>, Hugging Face, OpenRouter and Microsoft Azure AI Foundry, and are coming soon to Google Vertex AI Platform and other cloud service providers. NVIDIA Clara and Isaac GR00T are available on Hugging Face.<\/p>\n The models are available as NVIDIA NIM<\/a> microservices for secure, scalable deployment on DGX Cloud or any NVIDIA-accelerated infrastructure for maximum privacy and control.<\/p>\n
\nNemotron Nano 2 VL<\/a> provides document intelligence, image reasoning and video analysis.
\nNemotron Parse extracts text and tables from documents to take actionable insights.
\nNemotron Safety Guard<\/a> adds culturally aware multilingual moderation capabilities, detecting harmful content across 23 safety categories in nine languages.
\nNemotron retrieval-augmented generation (RAG)<\/a> models now feature advanced document extraction and unified retrieval across text, images, audio and video data sources.<\/p>\n
\nCadence<\/a> is accelerating chip designer productivity with its JedAI Platform and Nemotron open technologies.
\nCrowdStrike is building autonomous, continuously learning AI agents into its Agentic Security Platform with NVIDIA Nemotron.
\nPayPal is boosting the throughput and cost efficiency of its AI services by 50% with Nemotron.
\nSynopsys<\/a> is collaborating with NVIDIA to develop chip-design agents with the NVIDIA NeMo Agent Toolkit and Nemotron open models and data, used by NVIDIA engineers.
\nZoom<\/a> is utilizing Nemotron for customized agentic capabilities to align with the specific workflows of its customers.<\/p>\n
\nCosmos Transfer 2.5<\/a>: Produces higher-quality, photorealistic data from 3D scenes at one-third the size of Cosmos Transfer 1.
\nCosmos Reason<\/a>: A reasoning vision language model, now available as an NVIDIA NIM microservice, for advanced multimodal understanding.
\nCosmos Dataset Search<\/a>: Enables instant retrieval of training scenarios, cutting post-training cycles from months to days.
\nIsaac GR00T N1.6: Enhances reasoning, generalization and whole-body control for humanoid robots.<\/p>\n
\nClara La-Proteina: Creates 3D protein structures atom by atom at twice the length and complexity of previous models, enabling the design of better medicines, enzymes and materials.
\nClara Reason: A vision language model enabling chain-of-thought reasoning for radiology and medical imaging to advance explainable AI medical research.<\/p>\n