NVIDIA releases Nemotron 3 Super: throughput increases fivefold.

Nvidia launches Nemotron 3 Super, an open-source model with 1.2 trillion parameters. This model has been deeply optimized for the Blackwell architecture, with throughput in handling intelligent agent AI systems reportedly up to five times higher than the previous generation Nemotron Super model, and accuracy has doubled. The model aims to address the costly issues caused by "deep thought" and "context explosion," which have previously constrained the workflow speed of autonomous intelligent agents. It is currently supported on Google Cloud and Oracle, with support for Amazon AWS and Microsoft Azure coming soon.