Nvidia Opens the Nemotron Family Down to a 4GB Edge Model

Nvidia released Nemotron 3.5 Super VL, a 120-billion-parameter open model it claims delivers the fastest task completion and highest accuracy for specialized always-on agents. Alongside it came two edge-focused siblings: Nemotron 3.5 Bolt 74B-A9B at 24GB, tuned to balance intelligence against speed per gigabyte for local agentic work, and Nemotron 3.5 Bolt E4B, a 4GB model built for constrained devices. The lineup signals Nvidia's intent to own the inference layer at every scale, not just the data center.

Nemotron 3.5 Super VL is a 120B open-weights model aimed at always-on agent workloads

Nemotron 3.5 Bolt 74B-A9B fits in 24GB — viable on a single high-end workstation GPU

Nemotron 3.5 Bolt E4B squeezes into 4GB for genuinely edge and embedded deployments

The release extends Nvidia's open-model strategy from data center down to desktop and device

Source: https://blogs.nvidia.com/blog/nemotron-3-5-super-vl/