NVIDIA Unveils Nemotron 3.5 Super VL and Bolt for Local and Enterprise AI

NVIDIA has a new batch of open AI models for organizations and developers looking to run smarter AI agents. The company has introduced Nemotron 3.5 Super VL alongside two Nemotron Bolt models, targeting everything from enterprise data centers to local RTX-powered PCs.

NVIDIA Nemotron 3.5 Super VL promises faster AI agents

The highlight is Nemotron 3.5 Super VL, an open model featuring 120 billion parameters. NVIDIA claims it delivers up to five times faster throughput than the next-best models in its class, alongside strong accuracy across multimodal agentic tasks.

The model handles instruction following, tool use, multistep business workflows, and visual understanding. It can also process webpages, PDFs, charts, tables, and other structured documents, helping AI agents extract information and act on it.

Unlike closed AI models, Super VL offers open weights, datasets, and training recipes. Organizations can customize it using their own data, tools, and workflows, while keeping deployments on their own infrastructure. NVIDIA says Super VL is already being customized for financial analysis, cybersecurity, software development, and other enterprise applications. The model is available through Hugging Face, OpenRouter, ModelScope, and NVIDIA NIM.

Image credit: NVIDIA

Nemotron Bolt brings AI agents to local hardware

NVIDIA is also introducing two smaller models designed for edge and local AI workloads. Nemotron 3.5 Bolt 74B-A9B uses native 2-bit quantization to fit within 24GB of memory. NVIDIA targets RTX systems and DGX Spark, allowing users to run agentic tasks locally without sending workloads to the cloud.

The model supports tools such as llama.cpp, Ollama, LM Studio, Unsloth, and vLLM. NVIDIA claims it ranks first among open models in its class on PinchBench.

Meanwhile, Nemotron 3.5 Bolt E4B is designed for devices with a 4GB memory footprint. It targets conversational agents, local voice assistants, and interactive applications across GeForce RTX, Jetson, and RTX/DGX Spark hardware. As for the availability and support, Nemotron 3.5 Super VL and Bolt 74B-A9B are available now. Bolt E4B is scheduled to arrive on October 23.

Image credit: NVIDIA