Nvidia's New AI Models Run Locally on Your PC
This could let you run powerful AI without sending data to the cloud.
Nvidia introduced three Nemotron models. The 120-billion-parameter Nemotron 3.5 Super VL can analyze documents, interpret images, and use external tools to complete multistep tasks; it is available through Hugging Face, OpenRouter, ModelScope and Nvidia’s NIM deployment platform.
Nvidia also introduced two smaller Nemotron 3.5 Bolt models for local AI agents: Bolt 74B-A9B uses 2-bit quantization to fit within a 24GB memory footprint on compatible Nvidia RTX and DGX systems, while Bolt E4B requires 4GB and is scheduled for Oct. 23 on GeForce RTX, Jetson and RTX/DGX Spark hardware. The Bolt models could make locally run agents more practical, though privacy still depends on system configuration, application permissions and how connected tools handle data, and businesses still need to test accuracy, speed and resource use.
- Nvidia's new AI models can run on your own computer, not just the cloud.
- The smallest model needs only 4GB of memory, fitting on many devices.
- You can customize these models for specific tasks like reading documents or coding.
Why It Matters
This could lead to faster, more private AI on everyday devices, reducing cloud costs and data risks.