On August 11, Nvidia is developing a new generation of open-source artificial intelligence model, Nemotron 4, with an expected parameter scale of at least 1 trillion, aiming to compete with the world's leading open-source large models. Nvidia hopes to expand the application scope of AI through an open model ecosystem and further drive market demand for its GPU computing power. Nvidia has previously launched the Nemotron series of open-source models, including the Nemotron-4 340B series with 340 billion parameters, primarily used for large language model training data generation, model development, and enterprise AI applications.