Nvidia Announces New Open Models for Enterprise and Personal Computers

Nvidia announced new open-weight models developed for enterprise agents and edge computing systems on Thursday. In this context, the 120-billion parameter Nemotron 3.5 Super VL model and two lightweight Bolt models designed for local AI PCs and workstations were introduced.

Kari Briski, Nvidia's Vice President of Open Models and Software, stated that organizations can have an open and optimized model for their own layers without having to choose between capacity and deployability. The Super VL model is claimed to complete tasks faster than its competitors when post-trained on specialized tasks such as document intelligence, financial accounting, scientific experiments, medical insurance claims, and security attack resistance. The company announced that this model is used by prominent companies such as Skydance, EY, Ramp, Crowdstrike, Perplexity, and Salesforce.

On the other hand, two Bolt models—Nemotron 3.5 Bolt 74B-A9B 24GB and the smaller Bolt E4B—were launched to enable consumers to run their data locally without sending it to the cloud. While the 74-billion parameter Bolt model aims to strike a balance between speed and reasoning, the smallest model, E4B, was designed for chat agents. While Super VL and Bolt 74B-A9B models are available for immediate access, the Bolt E4B model is expected to be released later this month.