---
title: "NVIDIA Introduces Open Nemotron 3.5 Models for Autonomous AI Agents"
description: "NVIDIA has introduced three open Nemotron 3.5 models for autonomous AI agents. The central Super VL multimodal model with 120 billion parameters runs on local infrastructure and standard PCs."
date: 2026-10-16T14:36:00.837Z
lang: en
url: https://xab.info/en/posts/nvidia-introduces-nemotron-3-5-models
tags: [nvidia, nemotron, ai-agents, open-source-ai, llm]
author: "Alexander Buran"
publisher: "XAB.info"
---

# NVIDIA Introduces Open Nemotron 3.5 Models for Autonomous AI Agents

![Workstation setup with a monitor and a PC case displaying the green NVIDIA logo](https://xab.info/media/2026/10/16/nvidia-nemotron-3-5-modeli-dlya-ii-agentov/nvidia-nemotron-3-5-modeli-dlya-ii-agentov-1.webp)

## Key Points

- NVIDIA released three open Nemotron 3.5 models for autonomous AI agents
- The central Nemotron 3.5 Super VL model is multimodal, with 120 billion parameters (12 billion active)
- According to NVIDIA, the model is 5 times faster than class competitors and trained on 26 trillion tokens
- Models can be deployed on private infrastructure or a standard PC, with fine-tuning via NeMo

NVIDIA has introduced three new open models in the Nemotron 3.5 family, focused on running autonomous AI agents. According to the company, developers will be able to deploy them both on corporate infrastructure and on a standard personal computer. This was reported by RBC-Ukraine with reference to Interesting Engineering.

### Central Release: The Multimodal Super VL

The key new addition is the Nemotron 3.5 Super VL multimodal model, featuring 120 billion parameters. It is designed for visual data analysis, document processing, and executing multi-step workflows using external tools.

### Architecture: 120 Billion Parameters, 12 Billion Active

According to NVIDIA's official announcement, the model uses an architecture with a total of 120 billion parameters, of which 12 billion are active. This approach combines the capacity of a large neural network with a relatively low computational load per request, which is crucial for deployment on constrained hardware.

### Speed and Accuracy According to the Company

According to NVIDIA, the model performs tasks five times faster than competitors in its class. At the same time, in terms of analysis and reasoning capabilities, the neural network approaches the company's largest models while utilizing only a fraction of their parameters.

### Training and Fine-Tuning

The model was trained on 26 trillion tokens and supports fine-tuning via the NVIDIA NeMo framework, allowing companies to adapt it to their own tasks and data.

### Developer Accessibility

Special attention is given to accessibility: the open Nemotron 3.5 family models can be run on enterprise infrastructure or even a standard PC, lowering the barrier to entry for teams lacking access to cloud clusters.

## Fact-Check & Sources

- [В 5 раз быстрее аналогов: NVIDIA выпустила мощную модель ИИ для ПК](https://www.rbc.ua/ukr/news/5-raziv-shvidsha-analogi-nvidia-vipustila-1792155200.html) - Confirms the release of three open Nemotron 3.5 models, Super VL parameters (120B), 5x speed, training on 26 trillion tokens, fine-tuning via NeMo, and PC execution capability. Source — Interesting Engineering.

## FAQ

### What is Nemotron 3.5 Super VL?

It is a multimodal open model from the Nemotron 3.5 family with 120 billion parameters (12 billion active), designed for visual data analysis, document processing, and multi-step workflows using external tools.

### Where can Nemotron 3.5 models be run?

According to NVIDIA, they can be deployed on a company's own infrastructure or on a standard personal computer.

### How much faster is the model compared to competitors?

According to NVIDIA, the model performs tasks five times faster than competitors in its class; independent verification of this metric is not provided in sources.