Nvidia on Monday unveiled a new family of open-source artificial intelligence models that it said are faster, cheaper and more capable than its previous offerings, as open-source releases from Chinese AI laboratories continue to proliferate.

The U.S. chipmaker is best known for supplying the high-performance processors used by companies such as OpenAI to train proprietary, closed-source AI models that are then sold commercially. However, Nvidia has also been expanding its own portfolio of open-source AI software, offering models designed for applications ranging from physics simulations to autonomous driving.

The company said the latest models are intended to lower computing costs while improving performance, making them easier for researchers and enterprises to deploy. Nvidia's open-source models are already being integrated into commercial products, with companies such as Palantir Technologies incorporating them into their platforms.

The move underscores Nvidia's effort to strengthen its influence across the AI ecosystem, not only as the dominant hardware supplier but also as a provider of widely used software tools, at a time when global competition in open-source AI is intensifying.