Nvidia developing Nemotron 4 AI model family

Nvidia is preparing a new generation of open AI models that could push beyond the trillion-parameter mark, as the chipmaker seeks to strengthen its position in the rapidly evolving global open-model ecosystem.

Nvidia is reportedly working on a new family of artificial intelligence models called Nemotron 4, with the company aiming to compete with some of the leading open-source AI systems available globally.

The initiative comes as open and open-weight AI models gain increasing attention from developers and businesses. The trend has been driven partly by growing demand for more accessible AI technologies, while advances from Chinese developers have intensified competition with leading US-based AI companies.

Open models have also become part of a wider discussion around AI safety and cybersecurity. Recent incidents involving autonomous AI agents have highlighted concerns about how powerful models can be used when developers and users have greater freedom to modify or deploy them.

Nemotron 4 could cross the trillion-parameter mark

According to employees familiar with the project cited in the report, Nvidia's largest Nemotron 4 model is expected to feature at least 1 trillion parameters. Such a scale would place the model among the most ambitious open AI systems under development.

However, Nvidia has not committed to a launch date. The model is reportedly still undergoing development and has yet to complete its final training phase. People working on the project indicated that a release could potentially take place as early as late autumn, although the timeline remains subject to change.

Nvidia reportedly did not independently confirm the specific details reported about Nemotron 4. Instead, the company referred to its earlier statements about developing the model family.

Kari Briski, vice president of generative AI at Nvidia, said in an emailed statement, "Nvidia is ‌investing in Nemotron because we believe every company and every country needs ‌accessible frontier open models to strengthen safety and security, accelerate innovation, and provide ‌a foundation they can rely on from one generation to the next."

Nvidia expands its open AI portfolio

Nvidia's latest effort comes as the company continues to promote open AI development and broader industry collaboration around safety and security. Last month, it joined other technology companies in forming a coalition focused on creating and sharing tools for AI safety and cybersecurity.

The company has also backed open-weight AI models through an industry letter, arguing that broader access to such technologies can help maintain the pace of innovation.

Separately, Nvidia introduced Nemotron 3.5 Lightning, targeting practical enterprise and software-development applications. The model is designed for tasks such as code review, tool interaction, security monitoring and responding to billing-related queries.

Nvidia also unveiled NeMo Switchyard, an open-source model-routing library that can automatically determine which AI model is best suited to handle a particular task. Together, the developments underline Nvidia's broader strategy of supporting an ecosystem of open models and tools alongside its core AI hardware business.