NVIDIA develops Nemotron 4 model with trillion parameters

NVIDIA is developing a new generation of AI models called Nemotron 4, aiming to compete with the world's most advanced open-source models. According to project participants, the largest model in the series is expected to have at least one trillion parameters. The company plans to launch the model later this year, possibly in late autumn, though training is not yet complete. NVIDIA also released Nemotron 3.5 Lightning, optimized for code review, tool calls, and security alert monitoring, as well as the open-source NeMo Switchyard library for automatic task routing across models.