Nvidia Launches New Nemotron 3.5 AI Models

The chipmaker expands its open-weight portfolio with high-performance models designed for enterprise agentic workflows.

Updated on Oct. 15, 2026 in Artificial Intelligence

Bold flat-color editorial illustration of metallic server chassis with sharp geometric fins, evoking enterprise-grade computational infrastructure. — Nvidia has launched its Nemotron 3.5 AI model series, enabling enterprises to deploy high-performance agentic workflows on local infrastructure. AI Illustration. Upload story photo >

Live Poll

Do you believe the development of U.S.-based open-weight AI models will improve national security and technology?

Nvidia has released the Nemotron 3.5 Super VL, a 120-billion parameter AI model built to accelerate enterprise workflows. The company also introduced two compact edge models to support specialized business needs.

Why it matters

These models are designed to enable companies to handle high-volume agentic tasks locally, offering enterprises more control over their data infrastructure. By allowing post-training on internal datasets, Nvidia aims to improve task-specific precision for critical business processes.

The 120-billion parameter Nemotron 3.5 Super VL model delivers a 5x speed increase over competitors, while the smaller 74B and 4B edge models offer 24GB and 4GB memory capacities respectively. Post-training the base model yields a 44% accuracy gain in medical insurance claims.

The players

Nvidia

Nvidia is a leading American technology company that specializes in the design and manufacture of graphics processing units and artificial intelligence hardware.

The details

Enterprises can post-train these models on their own proprietary data and deploy them on local infrastructure to ensure privacy and efficiency. The rollout includes the Super VL flagship and two edge-focused versions, the Bolt 74B-A9B and Bolt E4B, to handle varying compute requirements.

Nvidia released the Nemotron 3.5 Super VL model on October 15, 2026.

The Tech Race

This release follows the industry shift toward open-weight models, mirroring the open-source strategies popularized by major research labs to capture market share. Nvidia is positioning its hardware and software ecosystem to remain central as enterprises move AI workloads from the cloud to local servers.

Business users may see improved speed and accuracy in automated backend tasks, such as processing medical insurance claims. By running these models on local hardware, companies can reduce reliance on external cloud providers while maintaining higher data security standards.

The takeaway

Enterprises looking to integrate AI should prioritize models that allow for local post-training to maximize accuracy for specific industry tasks. Adopting localized models can provide significant efficiency gains and better protection of proprietary information compared to off-the-shelf cloud solutions.