Nemotron

Nemotron
DeveloperNvidia
ReleaseNovember 15, 2023; 2 years ago (2023-11-15)
TypeLarge language model
LicenseVarious
Nvidia Open Model License
Websitedeveloper.nvidia.com/topics/ai/nemotron
Repositorygithub.com/NVIDIA-NeMo/Nemotron

Nemotron is a family of artificial intelligence models developed by Nvidia. It includes large language models and multimodal models intended for reasoning, computer programming, information retrieval and agentic AI applications. Nvidia has released open model weights, training data, software and training methods for parts of the family.

History

Nvidia introduced the Nemotron-3 8B models in November 2023 for enterprise generative artificial intelligence applications. Nemotron-4 340B followed in June 2024 and included base, instruction-tuned and reward models designed partly for generating synthetic data.

Nvidia later introduced Llama Nemotron, a series of reasoning models derived from Meta's Llama models. In December 2025, Nvidia announced the Nemotron 3 generation, beginning with Nemotron 3 Nano. Nemotron 3 Super and Ultra were released in 2026.

Models

Major model families
Family Released Purpose
Nemotron-3 2023 Enterprise applications
Nemotron-4 2024 Synthetic-data generation
Llama Nemotron 2025 Llama-based reasoning
Nemotron 3 2025 Agentic AI

The Nemotron 3 models use a hybrid architecture combining Mamba, Transformer, and mixture of experts components. So far, the number of active parameters is always 1/10 of total parameters.

Nemotron 3 models
Nemotron Model Released Parameters (total, in billions) Intended use
3 Nano 2025.12.15 30 Efficient agents
Super 2026.3.11 120 Agentic reasoning
Nano Omni 2026.4.28 30 Multimodal agents
Ultra 2026.6.4 550 Complex reasoning
3.5 Lightning 2026.8.11 30 Fast agents

Nano Omni supports text, image, video and audio input. Nemotron models can be downloaded for local deployment or accessed through Nvidia's application programming interfaces and NIM inference services.

Nemotron Coalition

Nvidia launched the Nemotron Coalition on March 16, 2026, at its GPU Technology Conference (GTC) as a collaboration between open model builders and AI developers. The first project is a base model, to be co-developed by Mistral AI and Nvidia for a future Nemotron 4 family (distinct from the 2024 release of Nemotron-4 340B). Mistral AI and Nvidia previously collaborated on Mistral NeMo 12B, an open-weight model trained on Nvidia DGX cloud and released in July 2024.

Nvidia is to contribute an undisclosed quantity of DGX cloud compute to train the base model. The eight founding members (with anticipated contributions according to Nvidia) were Black Forest Labs (multimodal capabilities), Cursor (coding benchmarks and evaluation datasets), LangChain (tool use and long-horizon reasoning), Mistral AI (customizable models), Perplexity AI (AI systems), Reflection AI (open systems), Sarvam AI (sovereign language AI development), and Thinking Machines Lab (data collaboration). In June 2026, Nvidia announced that four additional companies would join the coalition: H Company, Naver Cloud, Nous Research (maker of Hermes Agent), and Prime Intellect.

See also

References

  1. ↑ "NVIDIA Nemotron Open Model License". Nvidia. December 15, 2025. Retrieved July 29, 2026.
  2. 1 2 3 4 Lee, Jane Lanhee (December 15, 2025). "Nvidia unveils new open-source AI models amid boom in Chinese offerings". Reuters. Retrieved July 29, 2026.
  3. 1 2 Knight, Will (December 15, 2025). "Nvidia Becomes a Major Model Maker With Nemotron 3". Wired. Retrieved July 29, 2026.
  4. 1 2 "NVIDIA AI Foundation Models: Build Custom Enterprise Chatbots and Co-Pilots With Production-Ready LLMs". Nvidia Technical Blog. Nvidia. November 15, 2023. Retrieved July 29, 2026.
  5. 1 2 Nuñez, Michael (June 14, 2024). "Nvidia's Nemotron-4 340B model redefines synthetic data generation, rivals GPT-4". VentureBeat. Retrieved July 29, 2026.
  6. 1 2 Bercovich, Akhiad; Levy, Itay; Golan, Izik; Dabbah, Mohammad; et al. (May 2, 2025). "Llama-Nemotron: Efficient Reasoning Models". arXiv:2505.00949 [cs.CL].
  7. 1 2 Knight, Will (March 11, 2026). "Nvidia Will Spend $26 Billion to Build Open-Weight AI Models, Filings Show". Wired. Retrieved July 29, 2026.
  8. 1 2 "NVIDIA Nemotron 3 Ultra". Nvidia Research. Nvidia. June 4, 2026. Retrieved July 29, 2026.
  9. ↑ "Introducing Nemotron 3 Super: An Open Hybrid Mamba-Transformer MoE for Agentic Reasoning". Nvidia Technical Blog. Nvidia. March 11, 2026. Retrieved July 31, 2026.
  10. 1 2 "NVIDIA Nemotron 3 Nano Omni Powers Multimodal Agent Reasoning in a Single Efficient Open Model". Nvidia Technical Blog. Nvidia. April 28, 2026. Retrieved July 29, 2026.
  11. ↑ "What's New in Deep Learning & Artificial Intelligence from NVIDIA". NVIDIA. Retrieved 2026-10-10.
  12. ↑ Tarre, Mark. "NVIDIA unveils Nemotron 3.5 Lightning for local agents". IT Brief Australia. Retrieved 2026-10-10.
  13. ↑ Sirodot, Bertrand; Bronzati, Fabricio (November 7, 2025). "Introduction to NVIDIA Inference Microservices, aka NIM". Dell Technologies Info Hub. Dell Technologies. Retrieved July 29, 2026.
  14. ↑ "NVIDIA Nemotron". Nvidia. Nvidia. Retrieved July 29, 2026.
  15. 1 2 AI, Luke James Nemotron Coalition's first project is a base model co-developed with Mistral; Release, Open Sourced on (2026-03-16). "Nvidia's Nemotron coalition brings eight AI labs together to build open frontier models". Tom's Hardware. Retrieved 2026-10-11.
  16. 1 2 "NVIDIA Launches Nemotron Coalition of Leading Global AI Labs to Advance Open Frontier Models". NVIDIA Newsroom. Retrieved 2026-10-11.
  17. ↑ "Mistral NeMo | Mistral AI". Mistral. Retrieved 2026-10-11.
  18. ↑ "Mistral AI partners with NVIDIA to accelerate open frontier models | Mistral AI". Mistral. Retrieved 2026-10-11.
  19. ↑ Briski, Kari (2024-07-18). "Mistral AI and NVIDIA Unveil Mistral NeMo 12B, a Cutting-Edge Enterprise AI Model". NVIDIA Blog. Retrieved 2026-10-11.
  20. ↑ Moss, Sebastian (March 16, 2026). "Nvidia offers DGX Cloud to AI foundation model labs for new Nemotron Coalition: Locking in open source support". Data Centre Dynamics Ltd (DCD). Retrieved October 10, 2026.
  21. ↑ Lardinois, Frederic (2026-03-16). "Nvidia brings together AI labs to build the next generation of open base models". The New Stack. Retrieved 2026-10-11.
  22. ↑ Vert, Alyona (2026-03-18). "Nemotron 3 and the Surprising Coalition Building New AI in the Open". Turing Post. Retrieved 2026-10-11.
  23. ↑ Writers, NVIDIA (2026-05-21). "NVIDIA GTC Taipei at COMPUTEX: Live Updates on What's Next in AI". NVIDIA Blog. Retrieved 2026-10-11.