Open From Data Center to Desktop: NVIDIA Expands Nemotron Family With Highest Efficiency Super VL and Bolt Models
How-To How to actually use this
What changed: NVIDIA released three new open Nemotron 3.5 models—Super VL (120B), Bolt 74B-A9B (24GB), and Bolt E4B (4GB)—optimized for efficient agentic workloads from data center to desktop.
How to use it:
- Download the desired Nemotron 3.5 model (Super VL, Bolt 74B-A9B, or Bolt E4B) from NVIDIA’s official open model repository.
- Install required dependencies (e.g., TensorRT-LLM, CUDA) as specified in the model’s documentation.
- Load the model locally or in your agent framework using standard Hugging Face or NVIDIA AI Enterprise tools.
- Configure your always-on agent to use the model for low-latency, high-accuracy task completion.
- Monitor performance per GB (especially for Bolt variants) to optimize for your hardware constraints.
Good for: developers building efficient, always-on AI agents for edge or hybrid deployment.
NVIDIA today released Nemotron 3.5 Super VL, an open 120-billion-parameter model that delivers the fastest task completion and highest accuracy for specialized always-on agents. Joining it are two new edge models: Nemotron 3.5 Bolt 74B-A9B 24GB, optimized to balance intelligence and speed per GB for local agentic workloads, and Nemotron 3.5 Bolt E4B 4GB, optimized […]
Read original article on NVIDIA Blog →
