NVIDIA releases Nemotron 3 open-source AI model series, with throughput up to 4 times that of the previous generation.

NVIDIA has released the Nemotron 3 open-source AI model family, including Nano, Super, and Ultra versions. The Nano version (30 billion parameters) is already available, employing a hybrid expert architecture and claiming four times the throughput of its predecessor, a 60% reduction in inference token generation, and a 1 million token context window. This family is designed for multi-agent systems, aiming to address challenges such as communication overhead and inference costs.